What: Signature is an innovative, research-based HR conferenceWhen: Sunday, November 5 to Tuesday, November 7, 2023Where: The Hilton Orlando

Signature, the Premier HR Conference From McLean & Company, Returns in Five Days

Vice President Kamala Harris and Secretary of Commerce Gina Raimondo both spoke to the importance of the new institute, along with unveiling other safety measures the U.S. will deploy, on Wednesday at an A.I. Safety Summit in the United Kingdom.

The Biden Administration to Create New A.I. Safety Institute to Tackle A.I.'s 'Dangerous Capabilities'

Maury Alliance's final Breakfast with The Mayor featured city leaders discussing growth, planning and tourism's latest economic impact on Columbia.

Breakfast with The Mayor addresses growth, tourism and planning for the future

A court says Pennsylvania can't enforce a regulation to make power plant owners pay for their planet-warming greenhouse gas emissions.

Pennsylvania court permanently blocks effort to make power plants pay for greenhouse gas emissions

The public are to be consulted on plans that could ban new takeaways near schools or youth clubs in Bristol. The city council has approved a draft Local Plan that would also stop three or more takeaways from opening next door to each other. City Hall chiefs are concerned takeaways are contributing to rising levels of obesity.

New takeaways near city's schools could be banned

Plans to build 25 homes in a North Yorkshire village have been recommended for approval ahead of a planning meeting next week. North Yorkshire Council's Skipton and Ripon planning committee will meet on Tuesday to consider an application for the development in Cononley. A previous bid for the village was refused in December 2022.

Plan for village homes recommended for approval

Local law enforcement will implement heightened security measures for Sunday's New York City Marathon due to a rise in tensions around the globe, officials said on Wednesday. The top endurance athletes as well as the 50,

Athletics-NYPD plan heightened marathon security, citing global tensions

“It’s as if the sea star is completely missing a trunk, and is best described as just a head crawling along the seafloor,” study co-author and Stanford University evolutionary biologist Laurent Formery said in a statement. “It’s not at all what scientists have assumed about these animals.”

The sea star's whole body is a head

This suggests that, over the course of evolution, starfish and other echinoderms may have developed their five-sectioned bodies by losing this central trunk section, allowing them to hunt and feed differently from other animals.

Scientists Finally Solve Mystery of Where Starfish Keep Their Heads

New report outlines how caregivers face a series of financial and professional challenges, and ways financial advisers and employers can help.

Caregiving: A Growing Problem with Costly Impact on Budgets, Careers, Retirement

Plan A

plana

HackerNoon Reporter: Please tell us briefly about your background. I am the co-founder and CEO of Plan A and co-founder of the Greentech Alliance. The latter is a community of 1,000+ startups that are connected to over 350+ advisors from VC, media, and business. Prior to Plan A, I worked in investment banking, venture capital, and fintech in Asia and Europe. What's your startup called? And in a sentence or two, what does it do? We are Plan A. We have developed a science-based, one-stop SaaS platform that helps companies to decarbonize their operations. Our end-to-end software empowers businesses and organizations to accurately measure, monitor, reduce, and report their carbon footprint while improving their ESG performance. What is the origin story? A few years back, I went on a trip to Morocco, to explore the best places for surfing. This trip was definitely an eye-opening experience. I expected to see beautiful places to surf, but everything was really polluted, mostly covered in plastic. For the first time, my eyes were open to the negative impact we, humans, had on our planet. After this experience I got back to London and started to dig deeper into the topic of pollution and climate change, spending all my free time for the next 6 months educating myself on climate change and the correlation it had to pollution through online courses, books, and scientific papers. The knowledge I accumulated showed me that the situation worldwide was actually much worse than what was communicated across channels. My next step was to conduct a vast survey with 300 people and 100 companies to understand the behaviors and opinions about climate change. This was the final straw. After this, I was beyond convinced that addressing climate change needed a different approach. I quit my job and founded Plan A. What do you love about your team, and why are you the ones to solve this problem? The Plan A team consists of 40+ experts in climate data science, tech, business, design, and more. We gather some of the brightest minds such as our CTO and ML expert Oliver Spragg as well as our CEAO Vamsi Akuraju (ex Climate-KIC, WWF, Potsdam Institute). What I love most about our team is that we are so many like-minded people who share the same mission and vision day in and day out. We all want to tackle climate change by helping companies on their way to becoming truly sustainable. It’s this high level of cross-functional expertise combined with our passion and devotion to the cause that makes me confident we will be successful. If you weren’t building your startup, what would you be doing? There is so much evidence - and it’s ever-growing - that climate change is by far the biggest challenge of our time and that we have incredibly little time left to take action and turn the tide. That is why there would have been no alternative scenario for me - I would have founded and started building Plan A anyway. At the moment, how do you measure success? What are your core metrics? Within our company and throughout our teams we have various OKRs, KPIs, and metrics we want to deliver on to be successful. Our vision as a company, however, is that we want to empower organizations of all sizes to build a sustainable world. Therefore, two key metrics need to be mentioned here: The number of clients we can support in decarbonizing their activities across scopes 1, 2, and 3, as well as the number of carbon emissions they have reduced by using our SaaS platform. What’s most exciting about your traction to date? What is really exciting is that we have successfully supported more than 100 organizations from various industries around the world with our software and consultancy services to achieve their sustainability goals and reduce a significant amount of CO2 emissions. Through our information and education campaigns, we have reached, educated, and engaged over 1,000,000 people - from individuals to employees and directors to private, non-governmental, and governmental organizations. What technologies are you currently most excited about, and most worried about? And why? What really fascinates me is artificial intelligence and its many applications. I think AI is both a curse and a blessing because it can be used for good but also for the exact opposite. In the fight against climate change, AI is irreplaceable and a crucial part of our SaaS. What drew you to get published on HackerNoon? What do you like most about our platform? Hackernoon luckily found us. What I like most is the community, the valuable stories and insights, and the exchange. Vote for Plan A as the . startup of the year, Berlin

'Climate Change is By Far the Biggest Change of Our Time': Lubomila Jordanova, CEO Plan A

In terms of attracting new employees, startups have a tremendous edge. In contrast to large corporations, startups can provide employees with the opportunity to become highly wealthy by participating in the company's ownership. You may learn more about early-stage startups' equity distribution techniques by reading this article. Startup Equity Distribution Plan When it comes to the startup equity distribution plan, there are as many perspectives. Before signing any equity split agreement, it's usually a good idea to seek the opinion of professionals. Keep reading to know a few valuable pointers. What is Startup Equity? Ownership interest in a company is generally referred to as . Businesses usually use stock as a form of equity. While it is possible to own equity in any asset, this is not always the case. When you start a business, you should know how equity shares are classified and who receives what. startup equity Types of Startup Equity There are several ways a startup's owners might reward their employees and investors with premium pricing and other incentives based on their equity interest in the company. The common types of startup equity are listed below: Convertible Debt Notes A convertible debt note is a method of securing a loan from an investor by using the company's equity as collateral. The convertible note promises to either pay back the loan with interest or grant discounted equity during a subsequent financing cycle. In either case, investors receive their money back, plus interest, or a lesser price for a stake in the company. In order to obtain much-needed capital without having to give up any stock for a long time, this is a popular method for new enterprises. After a convertible loan note has been in place for 12 to 18 months, the company should undertake a Series A financing round in which investors can purchase discounted ownership in the company. Incentive Stock Option As a reward to employees, companies offer Incentive Stock Options (ISOs), which allow them to purchase shares of stock at a reduced price and avoid paying capital gains taxes. As ISO profits are treated as capital gains rather than regular income, they receive a tax advantage. Stock options that do not meet the Internal Revenue Code's standards to be referred to as ISOs are called Non-Qualified Stock Options (NSO). These profits are taxed as income rather than capital gains. How does Startup Equity Work? Employees and investors are typically given a stake in the company in exchange for their contributions. It is determined by criteria such as the timing of the equity distribution, the level of participation and commitment, and the company's current worth. Founders, on average, receive the greatest initial equity. Stakeholders deserve to have a share in their company's success, which is why startup equity is so important. Early investors also generally receive more equity than others who join the firm later because their investments are larger in relation to the company's initial valuation. It's also common for early employees to have a larger stake in the company than those who join later in the process. In addition to fundraising stages, equity distribution is tightly linked. Your financial situation inevitably shifts as you proceed through funding rounds, and in almost every case, how you allocate equity shifts as well. Things to Consider Before Splitting Startup Equity An equity split between the company's founders, investors, and employees can be a difficult task to do. The following is a list of things to keep in mind while determining the correct equity split for your new business. - Co-founders and employees may be ready to accept a lower wage because they feel their ownership part in the company will be worth more in the future. Among the best examples is Nike's logo designer, who was paid around $35 a share of the company's equity, which is now worth almost $640,000. Replacement of salary - As a thumb rule, the person who came up with the company's primary value proposition is entitled to the most stock. However, this isn't always the case. The value of a good idea does not outweigh the value of actual, physical contributions of capital and sweat equity. The idea to split equity between two or more co-founders should be based on earlier developmental efforts contributed by each of them. As an example, one of the co-founders of Instagram ended up obtaining a 40% stock position because he was the creator of a company. Another co-founder came on board later and was given a 10% ownership in the company, and the balance was dispersed among investors and staff. A reasonable equity split - The stages of a startup determine the share percentage. It is customary for co-founders and early employees of a startup to receive a higher share of the company's ownership (Before series A funding, seed round) as compensation for their work and risk-taking in the company's early stages. Stage of a startup - A vesting schedule is generally recommended regardless of how equity is allocated among co-founders. The equity split agreement specifies in detail when and how co-founders might exercise their stock option rights under a vesting schedule. An effective vesting schedule can help prevent a startup firm from being collapsed by the departure of one of its co-founders, who may take a significant chunk of the company's worth with them. Vesting schedules How to effectively split equity in a startup? In the early stages of a new company, the founders possess all of the company's equity. As long as you are the only founder, you have full ownership of the business. More individuals investing their time and money means more likely you will have to divide up the company's stock and give some of it away to those who have supported you. In some cases, you may want to grant a non-founder or financial backer an equity part in your company. A common example is when employees are compensated with equity rather than wages or salaries. The below brief explains how to split equity effectively. Founders Entrepreneurs who start a company with two or more co-founders face the challenge of dividing startup equity by role fairly and evenly. At this early stage, most likely, you still hold all of the company's equity. Splitting this among the founders before handing a stake to investors is a good idea. Each founder should receive a fair share even if their investment does not match that of the other co-founders in the ideal situation. When beginning a new firm, it's not only about how much money each founding partner contributes. Other factors to consider include, but are not limited to: Each entrepreneur adds a unique set of skills and knowledge to the company. One partner's contribution of tangible assets or equipment The amount of time that each party devotes to the startup venture It's reasonable for entrepreneurs to demand compensation for all of these variables. Investors In order to award equity to third-party investors, you need to have a firm foundation on how to distribute equity to founders who have reached an agreement on its division. The amount of equity you're ready to sell must be determined at every stage of investment, from the seed round through Series A, B, and C. An equity stake in your company is exchanged for money from investors. Again, the quantity of stock received by each investor should reflect the amount invested. Well-known investors may try to cut the price or increase their equity stake because of their experience, unique knowledge, and/or success in the past. Accepting a cheaper offer is up to the majority of stakeholders; however, consider how other investors may feel if they discover you have paid more for their part of the company. Whenever you transfer equity to a new outside investor, consider how much of your original investment is being lost. For example, ensure that the company's founders have at least a 51% ownership in the company's equity. Employees Employee stock and vesting schedules might well be necessary at some point in a new company's life cycle. Typically, this is defined as: The primary founders of the company. First-time hires (the first 25 hires by the startup) Soon-to-be-hired staffers (the 26th and subsequent individuals to be hired) They can have a huge impact on your company's success because they are the first ones to walk in. There is a chance that they can bring unique expertise or access to markets that you otherwise would not have. However, you may not be able to afford to pay your first staff a salary due to a lack of established income or cash flow. You can avoid this problem if you give your first employees a stake in your company's long-term success through a stock offer. A common technique to figure this out is to compare the current value of your firm to the value it will hypothetically be worth when a new hire joins you. This increased value determines how much equity you can provide to that person and yet break even on hiring them. Even though you give away some stock, you can still make a profit if you give them less than the break-even point. However, an equity share provides them with a genuine interest in ensuring that your company succeeds to the fullest potential extent. Conclusion We hope you found this post helpful in figuring out how to split your shares in a way that doesn't insult anyone and doesn't get misplaced. 90% of the time, individuals agree on how the company's future profits will be split, preventing any potential issues as startups grow swiftly and unpredictably.

How to Effectively Plan a Startup Equity Distribution?

Plan A didn’t work but maybe a Plan B can? This really is a topic that I have been hit with by many individuals and also one that has been running through my head lately which I’d like to share. In whatever you just have a backup plan you “Plan A” might not be much more dominant then your “Plan B”. Before we get started for those who don’t know me and for those who never reached to the bottom of my article for whatever reason: Here is a little background info about me so you can have a general knowledge of what I’m about to say and about what I really do. I’m 16 years old | Programmer |CEO of Documented Press| Content Marketer If you’re interested in following up with me check my site: https://jdombele.wixsite.com/jayson If you’re interested in my PitchUs startup please check this link below: https://jdombele.wixsite.com/pitchus Want news about politics,business,tech and more? https://documentedpress.wordpress.com/ If that is not enough information about me — well thank you for having the interest in getting to know me because today I will share much about me. I’m a junior in a Chicago high-school striving to launch a startup that would dent the world and leave a mark of positivism. While I’m working on my startup I previously launched Documented Press (if you yet don’t know — it’s a news media organization which allows you to hear from voices people find “unknown” or “unworthy”) which was launched January 31,2016. Documented Press itself has attracted 18,000+ viewers over the last two years. I am looking forward to pursuing being a programmer (well, like I already am but I want to be a known programmer or have it like my job) and I also want to work in the sector. business I sometimes say I already am in the business sector even tho I run a non-profit organization but my name doesn’t really surface people minds and that’s what I’m looking forward to changing. But, I also need to remember that I’m a student and school is also important. There was so much bias I encountered when I was entering the programming sector about how everyone just codes all day and not really do much and business owners tend to work so long — But, That’s not really going to lead you to success. As a matter of fact, that was one problem that led me to juggle my grades at the very end of the semester. By the way if you looking forward to starting up a company and you don’t have the materials or the “official guide” to starting up. Go to PitchUs official page and subscribe to my newsletter so you can get up to date for app release. https://jdombele.wixsite.com/pitchus Lately, thoughts been running through my head such as “Which university I should attend?”, “How can I transfer college credits to my potential university?” , “How high do I want my SAT score?” and much more other questions. And I been sitting down to really think about how I should plan that out. But, I told myself the matter of success only comes when you giving full effort even to the backup plan! This is how I personally planned everything for my life for I can say 1 & 1/2 year: Start writing Medium articles about things I learned, my experiences, advice, and pitch in my . startup As I move forward I will be working on my app PitchUs (which was previously known as ShareME). On the side I will be focusing in school and keeping my grades up. Now if PitchUs works I will collaborate all my articles which will include all the values I used to make my startup successful. And, if PitchUs fail I will still make a book on what I learned about creating a startup and reasons why I failed and how to prevent the same mistakes. Assuming that PitchUs failed I can use my university as a backup to bring in a higher profile in education to get a job, get more experienced and work on another startup in after hours. (Shout out to my dad for that idea). In the long run, I want to build my audience on every platform I use regardless of any situation because when my startup is released I have a higher chance to gather high levels of traction. One key you need to put in mind is this: Don’t make obvious mistakes Understand that having a backup is a smart way of playing the game. You never know that backup plan can whine up being your main plan throughout execution. Thank you for reading my article! Feel free to leave a comment or learn much more about me below!! If you missed my last article: Out-execute & adapt or you set to fail More about Myself: I’m 16 years old | Programmer |CEO of Documented Press| Content Marketer If you interested in following up with me check my site: https://jdombele.wixsite.com/jayson If you interested in my PitchUs startup please check this link below: https://jdombele.wixsite.com/pitchus Want news about politics,business,tech and more? https://documentedpress.wordpress.com/ // I am currently looking for an editor for my article so if you interested please shoot me an email! jdapple4357@gmail.com // Follow Me: | TWITTER LINKEDIN | YOUTUBE | INSTAGRAM Any questions, business inquiries or concerns please feel free to email me @: jdapple4357@gmail.com

Don’t despise a Plan B because it can become your Plan A

The global number of businesses bought and sold in 2014 was than the year before. As entrepreneurs that started their companies in the boom of the 70s and 80s are starting to retire, this number is set to grow even more in 2015 in 2016. 11% higher Does that mean that it is difficult to sell your in this period? Probably not. The amount of capital in the form of equity is rising and there are plenty of parties that are acquiring. However, the competition is a little higher and, to get the best deal, it is better to prepare for it. business So how do you achieve a successful exit? and what does it mean anyway? An exit is when you manage to show the right value for what you have done — reflected in your company. This does not mean oversell, this means make the company and present it at it’s full potential. Indeed, if you don’t do this, you are underselling, allowing a newcomer to take over the company, make those little changes and maybe even resell it straight away. successful So there are actions that can be taken beforehand that can prepare the company to the sale and can allow you to bring it to the highest potential. Of course, the longer you prepare for an exit, the more actions you will be able to undertake. [That is why you should think about exit from the very beginning… but that is for another post ] In general, there are shorter term actions that can be taken even some months in advance. Create a company that succeeds without you The main thing, which is sometimes overlooked, is to create a company that can function and succeed even when you are taken out of the picture. This is easier said than done, and a lot of entrepreneurs underestimate how much of the company success is tied to them personally. They sometimes hold all the relationships, they have created a reputation of trustworthiness for themselves (that could not be reflected on the company as a whole), and sometimes they just do most of the jobs that nobody else can do. If you are about to remove yourself from the company, it is better to iron out these things first. Make sure to pass your connections to somebody who is going to stay, make sure that your knowledge is transferred, and put people in a position to thrive in doing what you did, maybe even more than yourself. Invest in your brand Buyers of your company will not know everything you know about your business. To attract the best ones and to properly present the maximum value of your company, a simple action that you can undertake is invest in your brand. Again, if you remove yourself as a founder, and the company becomes a stand-alone, it will not be able to rely on your reputation. It has to have a stand-alone reputation, and that is it’s brand. Quick actions that can be taken are, for example, renaming it if the name was tied with yours, start communicating at company level instead of personal level, like ‘We at company X do this’ instead of ‘I do this’ and start reinforcing this separation within the company. Increase efficiency Do you know how most of the Private Equity Funds make their money? They buy into companies, take actions to improve efficiency and then sell their participation at a premium, usually a hefty premium. So why not take some lessons from them and take those actions before somebody else takes them? Of course, efficiency can never be 100%, but sometimes, especially as a business matures, certain things get looked over and grow into easy-to-eliminate inefficiencies. Slow computers, moving the factory closer to the marketing office, replacing some costly but outdated machinery; These things are usually delayed actions because they do not bring immediate revenues and they are costly. However, they are going to bring the company closer to it’s full potential, thus increasing the likelihood of a potential exit. Track value improvements All these actions are fine, and intuitively they lead to a better exit outcome. They are also good general practices if you are just running the company and you have not thought about exits so far. However, it is much easier to improve things if you can track this improvement. Generally, especially in financial terms, you can look at indicators for efficiency. Do marketing surveys for branding and give different titles to your key employees to make sure that the company can function without you. But how do you know if these actions are ‘working’? All these actions are going to reflect on one and only metric, how much your company is worth. All these actions are going to increase how much the company is worth and thus bring the transaction value closer to its full potential. So it’s important to keep track of it, and make sure that these actions are actually reflected in a value increase. You can use different ways to track company value, you can create your own spreadsheet, talk to consultants, or (of course) try out Equidam. Remember, the best way to make the exit more successful is to make the company more successful and these are just some of the actions that can be undertaken to grow your value, bring it closer to its full potential, and get an exit as successful as possible for all the hard work that went into building the company. Originally published at www.equidam.com on May 15, 2015.

How to Plan a Successful Exit for your Business

Table of links ABSTRACT 1 INTRODUCTION 2 BACKGROUND: OMNIDIRECTIONAL 3D OBJECT DETECTION 3 PRELIMINARY EXPERIMENT 3.1 Experiment Setup 3.2 Observations 3.3 Summary and Challenges 4 OVERVIEW OF PANOPTICUS 5 MULTI-BRANCH OMNIDIRECTIONAL 3D OBJECT DETECTION 5.1 Model Design 6 SPATIAL-ADAPTIVE EXECUTION 6.1 Performance Prediction 5.2 Model Adaptation 6.2 Execution Scheduling 7 IMPLEMENTATION 8 EVALUATION 8.1 Testbed and Dataset 8.2 Experiment Setup 8.3 Performance 8.4 Robustness 8.5 Component Analysis 8.6 Overhead 9 RELATED WORK 10 DISCUSSION AND FUTURE WORK 11 CONCLUSION AND REFERENCES 3 PRELIMINARY EXPERIMENT We present in-depth observations on methods to improve the accuracy of the baseline BEVDet model mentioned in Section 2. Based on the results, we derived key insights and challenges in the design of Panopticus. 3.1 Experiment Setup Based on the baseline BEVDet model, we established 16 variants with different detection capabilities by combining the design parameters in Table 1. Each variant utilizes one of four combinations of ResNet backbones and input resolutions listed in Table 1, such as ResNet34 backbone with a resolution of 256×448, which we refer to as BEVDet-R34. Also, each variant incorporates one of two DepthNets, differing in depth prediction capabilities. Last, a group of variants utilizes a temporal fusion technique, further enhancing BEV feature generation. To compare with the state-of-the-art LiDARbased method, we profiled the CenterPoint [54] model. For model training, we utilized the nuScenes dataset [2] with ground-truth annotated 850 scenes, each of which contains a 20-second duration of video frames. Every frame provides 360° horizontal FOV through six camera images, each with a resolution of 900×1,600. For evaluation metrics, we used average precision (AP), widely employed to assess detection accuracy, and true positive (TP) errors. TP errors include average translation error (ATE) in meters and average velocity error (AVE) in meters per second, which represent the prediction errors of 3D location and speed for correctly detected objects, respectively. All metrics are calculated for each object type, such as cars and bicycles, and averaged across all types to report the model’s final detection performance, i.e., mAP, mATE, and mAVE. Inference latency was profiled on a recent edge device, Jetson AGX Orin [8] with a CUDA [6] GPU. 3.2 Observations Accuracy-latency trade-offs in different models. Both detection accuracy and latency are crucial factors in building a 3D detection system. As shown in Figure 3, the profiled 3D detection models exhibit various accuracy-latency tradeoffs. Increasing both backbone capacity and input resolution clearly enhances mAP, but leads to high latency. For example, BEVDet-R152, which has the largest backbone capacity and resolution, achieves a 1.6× mAP improvement compared to BEVDet-R34. However, the inference latency is substantially increased by +717ms. Incorporating a dense DepthNet and temporal feature fusion into BEVDet-R152 further improves mAP by 12.1% and 16.1%, respectively, outperforming LiDARbased CenterPoint. However, employing all of these methods to enhance camera-based detection can result in considerable inference latency on resource-constrained edge devices. Detection performance over various object properties. Real-world outdoor scenes encounter objects with diverse characteristics such as moving speed. We categorized objects based on the property and its levels defined in Table 2. For brevity, we focused on the distance between objects and the camera, as well as object velocity. As shown in the first row of Figure 4, all models achieve high mAP for objects with a D0 distance level. However, as the object distance level increases, lightweight models show significantly lower mAP. For example, the mAP of BEVDet-R34 for D3 objects is 2.7× lower than that of BEVDet-R152. The increase in errors of location and velocity predictions for distant and fast-moving objects, as shown in the second row of Figure 4, is mitigated in heavier models. Specifically, leveraging both temporal fusion and dense DepthNet reduces velocity prediction error for V3 objects by 2.4× compared to BEVDet-R50. Reducing the prediction errors of object properties is important for safety-critical applications. For instance, autonomous robots plan a navigation path based on the understanding of surrounding objects’ positions and movements. In summary, the detection capabilities of models vary across diverse object properties, requiring the use of appropriate models with an awareness of objects’ characteristics Detection performance over diverse spatial distributions. We examined how detection performance varies depending on spatial characteristics. Figure 5 shows the spatial object distribution and mAP in three example scenes from the nuScenes dataset. For each camera view (e.g., CAM1), we Figure 5: Distribution and mAP in example scenes. analyzed the spatial distribution by calculating both the number of objects and their average distance from the camera. Figure 5a shows that the overall scene complexity increases from Scene A to C, considering distribution in all camera views. Therefore, as shown in Figure 5b, the overall mAP gradually decreases. Note that the relative difference in distribution and mAP is even more significant across camera views. In Scene A, CAM4, which captures many distant objects, has the lowest mAP, whereas camera views with a smaller number of closer objects, such as CAM5, have a higher mAP. Meanwhile, in Scene C, the mAP between CAM3 and CAM5 differs by as much as 7.1×, indicating significant differences in spatial complexities. Consequently, diversity in spatial distribution across scenes and camera views necessitates model selection for each view to achieve the best performance. 3.3 Summary and Challenges Based on the observations, we derived insights to enhance both the accuracy and efficiency of omnidirectional 3D detection. Prior BEV-based models have made advancements in camera-based detection accuracy, but may result in substantial computational demands for a resource-constrained device. We observed that the models’ detection capabilities differ by object properties such as distance and velocity. Moreover, the variance in spatial distribution across scenes, even among different camera views within the same scene, suggests a need for employing diverse models to enhance overall performance. The experimental results inspired us to explore a high-accuracy omnidirectional perception scheme tailored for limited edge resources. This scheme would utilize the appropriate models for multi-view images, considering the models’ capabilities and the spatial distribution variations across camera views However, realizing this scheme poses several challenges. First, previous BEV-based models lack dynamic inference capabilities to process each camera view with varying inference configurations, such as the choice of backbone or DepthNet types. An approach to load all possible models into memory and assign them across views [28] may incur significant memory and latency overheads due to the redundant components among models. Moreover, the architecture of the BEV-based model needs to be adjustable considering various constraints such as device capabilities and latency requirements. Second, to maximize detection accuracy while meeting the latency requirements, the optimal inference configurations for all camera views must be established at each video frame. This requires real-time analysis of (1) changes in the surrounding spatial distribution within a dynamic environment and (2) the expected performance of different inference configurations. Authors:

Jeho Lee
Chanyoung Jung
Jiwon Kim
Hojung Cha This paper is available on arxiv under CC by 4.0 Deed (Attribution 4.0 International) license. Table of links Table of links ABSTRACT ABSTRACT ABSTRACT 1 INTRODUCTION 1 INTRODUCTION 1 INTRODUCTION 2 BACKGROUND: OMNIDIRECTIONAL 3D OBJECT DETECTION 2 BACKGROUND: OMNIDIRECTIONAL 3D OBJECT DETECTION 2 BACKGROUND: OMNIDIRECTIONAL 3D OBJECT DETECTION 3 PRELIMINARY EXPERIMENT 3 PRELIMINARY EXPERIMENT 3 PRELIMINARY EXPERIMENT 3.1 Experiment Setup 3.2 Observations 3.3 Summary and Challenges 4 OVERVIEW OF PANOPTICUS 4 OVERVIEW OF PANOPTICUS 4 OVERVIEW OF PANOPTICUS 5 MULTI-BRANCH OMNIDIRECTIONAL 3D OBJECT DETECTION 5 MULTI-BRANCH OMNIDIRECTIONAL 3D OBJECT DETECTION 5 MULTI-BRANCH OMNIDIRECTIONAL 3D OBJECT DETECTION 5.1 Model Design 6 SPATIAL-ADAPTIVE EXECUTION 6 SPATIAL-ADAPTIVE EXECUTION 6 SPATIAL-ADAPTIVE EXECUTION 6.1 Performance Prediction 5.2 Model Adaptation 6.2 Execution Scheduling 7 IMPLEMENTATION 7 IMPLEMENTATION 7 IMPLEMENTATION 8 EVALUATION 8 EVALUATION 8 EVALUATION 8.1 Testbed and Dataset 8.2 Experiment Setup 8.3 Performance 8.4 Robustness 8.5 Component Analysis 8.6 Overhead 9 RELATED WORK 9 RELATED WORK 9 RELATED WORK 10 DISCUSSION AND FUTURE WORK 10 DISCUSSION AND FUTURE WORK 10 DISCUSSION AND FUTURE WORK 11 CONCLUSION AND REFERENCES 11 CONCLUSION AND REFERENCES 11 CONCLUSION AND REFERENCES 3 PRELIMINARY EXPERIMENT 3 PRELIMINARY EXPERIMENT We present in-depth observations on methods to improve the accuracy of the baseline BEVDet model mentioned in Section 2. Based on the results, we derived key insights and challenges in the design of Panopticus. 3.1 Experiment Setup 3.1 Experiment Setup Based on the baseline BEVDet model, we established 16 variants with different detection capabilities by combining the design parameters in Table 1. Each variant utilizes one of four combinations of ResNet backbones and input resolutions listed in Table 1, such as ResNet34 backbone with a resolution of 256×448, which we refer to as BEVDet-R34. Also, each variant incorporates one of two DepthNets, differing in depth prediction capabilities. Last, a group of variants utilizes a temporal fusion technique, further enhancing BEV feature generation. To compare with the state-of-the-art LiDARbased method, we profiled the CenterPoint [54] model. For model training, we utilized the nuScenes dataset [2] with ground-truth annotated 850 scenes, each of which contains a 20-second duration of video frames. Every frame provides 360° horizontal FOV through six camera images, each with a resolution of 900×1,600. For evaluation metrics, we used average precision (AP), widely employed to assess detection accuracy, and true positive (TP) errors. TP errors include average translation error (ATE) in meters and average velocity error (AVE) in meters per second, which represent the prediction errors of 3D location and speed for correctly detected objects, respectively. All metrics are calculated for each object type, such as cars and bicycles, and averaged across all types to report the model’s final detection performance, i.e., mAP, mATE, and mAVE. Inference latency was profiled on a recent edge device, Jetson AGX Orin [8] with a CUDA [6] GPU. 3.2 Observations 3.2 Observations Accuracy-latency trade-offs in different models. Both detection accuracy and latency are crucial factors in building a 3D detection system. As shown in Figure 3, the profiled 3D detection models exhibit various accuracy-latency tradeoffs. Increasing both backbone capacity and input resolution clearly enhances mAP, but leads to high latency. For example, BEVDet-R152, which has the largest backbone capacity and resolution, achieves a 1.6× mAP improvement compared to BEVDet-R34. However, the inference latency is substantially increased by +717ms. Incorporating a dense DepthNet and temporal feature fusion into BEVDet-R152 further improves mAP by 12.1% and 16.1%, respectively, outperforming LiDARbased CenterPoint. However, employing all of these methods to enhance camera-based detection can result in considerable inference latency on resource-constrained edge devices. Detection performance over various object properties. Real-world outdoor scenes encounter objects with diverse characteristics such as moving speed. We categorized objects based on the property and its levels defined in Table 2. For brevity, we focused on the distance between objects and the camera, as well as object velocity. As shown in the first row of Figure 4, all models achieve high mAP for objects with a D0 distance level. However, as the object distance level increases, lightweight models show significantly lower mAP. For example, the mAP of BEVDet-R34 for D3 objects is 2.7× lower than that of BEVDet-R152. The increase in errors of location and velocity predictions for distant and fast-moving objects, as shown in the second row of Figure 4, is mitigated in heavier models. Specifically, leveraging both temporal fusion and dense DepthNet reduces velocity prediction error for V3 objects by 2.4× compared to BEVDet-R50. Reducing the prediction errors of object properties is important for safety-critical applications. For instance, autonomous robots plan a navigation path based on the understanding of surrounding objects’ positions and movements. In summary, the detection capabilities of models vary across diverse object properties, requiring the use of appropriate models with an awareness of objects’ characteristics Detection performance over various object properties. Detection performance over diverse spatial distributions. We examined how detection performance varies depending on spatial characteristics. Figure 5 shows the spatial object distribution and mAP in three example scenes from the nuScenes dataset. For each camera view (e.g., CAM1), we Detection performance over diverse spatial distributions. Figure 5: Distribution and mAP in example scenes. Figure 5: Distribution and mAP in example scenes. analyzed the spatial distribution by calculating both the number of objects and their average distance from the camera. Figure 5a shows that the overall scene complexity increases from Scene A to C, considering distribution in all camera views. Therefore, as shown in Figure 5b, the overall mAP gradually decreases. Note that the relative difference in distribution and mAP is even more significant across camera views. In Scene A, CAM4, which captures many distant objects, has the lowest mAP, whereas camera views with a smaller number of closer objects, such as CAM5, have a higher mAP. Meanwhile, in Scene C, the mAP between CAM3 and CAM5 differs by as much as 7.1×, indicating significant differences in spatial complexities. Consequently, diversity in spatial distribution across scenes and camera views necessitates model selection for each view to achieve the best performance. 3.3 Summary and Challenges 3.3 Summary and Challenges Based on the observations, we derived insights to enhance both the accuracy and efficiency of omnidirectional 3D detection. Prior BEV-based models have made advancements in camera-based detection accuracy, but may result in substantial computational demands for a resource-constrained device. We observed that the models’ detection capabilities differ by object properties such as distance and velocity. Moreover, the variance in spatial distribution across scenes, even among different camera views within the same scene, suggests a need for employing diverse models to enhance overall performance. The experimental results inspired us to explore a high-accuracy omnidirectional perception scheme tailored for limited edge resources. This scheme would utilize the appropriate models for multi-view images, considering the models’ capabilities and the spatial distribution variations across camera views However, realizing this scheme poses several challenges. First, previous BEV-based models lack dynamic inference capabilities to process each camera view with varying inference configurations, such as the choice of backbone or DepthNet types. An approach to load all possible models into memory and assign them across views [28] may incur significant memory and latency overheads due to the redundant components among models. Moreover, the architecture of the BEV-based model needs to be adjustable considering various constraints such as device capabilities and latency requirements. Second, to maximize detection accuracy while meeting the latency requirements, the optimal inference configurations for all camera views must be established at each video frame. This requires real-time analysis of (1) changes in the surrounding spatial distribution within a dynamic environment and (2) the expected performance of different inference configurations. Authors: Jeho Lee
Chanyoung Jung
Jiwon Kim
Hojung Cha Authors: Jeho Lee Chanyoung Jung Jiwon Kim Hojung Cha Jeho Lee Jeho Lee Chanyoung Jung Chanyoung Jung Jiwon Kim Jiwon Kim Hojung Cha Hojung Cha This paper is available on arxiv under CC by 4.0 Deed (Attribution 4.0 International) license. This paper is available on arxiv under CC by 4.0 Deed (Attribution 4.0 International) license. available on arxiv available on arxiv

Optimizing Omnidirectional 3D Object Detection for Edge AI

A New Architecture With Flexible QoS Control for Virtual Desktop Infrastructure

When MicroStrategy's stock plummeted 55% from its November high, most focused on the price drop itself. But they missed something far more important: A potential $8.2 billion debt bomb that could force the liquidation of nearly 500,000 Bitcoin. Here's the uncomfortable truth: MicroStrategy created the most aggressive leveraged Bitcoin position in history, and now the bill might be coming due. Think about it. Right now: MicroStrategy holds 499,096 Bitcoin worth $43.7 billion
They've borrowed $8.2 billion to acquire it
Their stock has crashed 55% in three months
Their debt covenants have a "fundamental change" clause But here's what everyone misses: This isn't just about Bitcoin. It's about financial engineering that could collapse. The Great Bitcoin Gamble Let's break down what's actually happening: MicroStrategy's entire business model relies on a flywheel effect: Borrow money through 0% convertible notes
Buy Bitcoin and drive price higher
Sell new shares at premium and buy more Bitcoin
Repeat But here's where it gets interesting: This flywheel only works in one direction. The Debt Structure Consider MicroStrategy's debt: $8.2 billion in convertible notes
Most maturity dates between 2027-2032
$3 billion due in 2029 alone
Conversion prices ranging from $39.80 to $672.40 These convertible notes contain a critical trigger – a "fundamental change" clause that could force early redemption if shareholders approve "any plan or proposal for the liquidation or dissolution of the Company." The Technical Trigger Here's the billion-dollar insight: If Bitcoin's value falls significantly, MicroStrategy's assets could slip below its liabilities, creating a perverse incentive for shareholders. Think about it: If Bitcoin falls 50%+ and stays down
Convertible notes can't be converted (underwater)
Debt comes due starting in 2027
Shareholders might prefer liquidation to pay themselves first The Michael Saylor Factor Saylor currently holds 46.8% of voting power, making a liquidation vote difficult. He insists they wouldn't sell "even if Bitcoin fell to $1." But this ignores market realities: Creditors have legal rights
Stock can't raise capital if it collapses
The flywheel needs constant new money
Cash flow can't cover interest payments The Market Verdict MicroStrategy's stock has experienced wild swings: Up 500% in the Bitcoin bull run
Down 55% in the recent correction
Now trading around $252, down from $560+ in November While the company's average Bitcoin purchase price is $66,357, their ability to service debt doesn't depend on being profitable – it depends on continued capital market access. The Real Questions The true vulnerabilities aren't about forced liquidation today. They're about: Can MicroStrategy refinance debt in a Bitcoin bear market?
Will investors keep buying shares if Bitcoin falls further?
What happens when convertible notes mature underwater?
How sustainable is a "never sell" strategy with massive debt? The Bitcoin Cult Factor As we noted in our analysis of Saylor's approach, he's positioned Bitcoin not as currency but as "digital gold" – a philosophical stance that's created cult-like devotion. The investment thesis relies on a permanent, upward trajectory. "If Bitcoin reaches $3M it would mean that something really bad is going on out there." But the flip side is troubling: What if it falls significantly and stays down? The Potential Endgame Three scenarios emerge: Bitcoin rises enough to make all convertible notes profitable, solving the problem
Bitcoin trades sideways, creating refinancing challenges when debt matures
Bitcoin falls significantly, potentially triggering a liquidation decision The Ultimate Irony Here's the greatest paradox: While Saylor positions Bitcoin as a hedge against monetary instability, he's built perhaps the most leveraged, fragile Bitcoin position possible. Welcome to the great MicroStrategy experiment. The question isn't whether Bitcoin has long-term value. The question is: Can a public company hold billions of it through debt without ever selling? The flywheel that took MicroStrategy to the moon works in both directions. And gravity is a powerful force. When MicroStrategy's stock plummeted 55% from its November high, most focused on the price drop itself. But they missed something far more important: A potential $8.2 billion debt bomb that could force the liquidation of nearly 500,000 Bitcoin. But they missed something far more important: A potential $8.2 billion debt bomb that could force the liquidation of nearly 500,000 Bitcoin. Here's the uncomfortable truth: MicroStrategy created the most aggressive leveraged Bitcoin position in history, and now the bill might be coming due. MicroStrategy created the most aggressive leveraged Bitcoin position in history, and now the bill might be coming due. Think about it. Right now: MicroStrategy holds 499,096 Bitcoin worth $43.7 billion They've borrowed $8.2 billion to acquire it Their stock has crashed 55% in three months Their debt covenants have a "fundamental change" clause MicroStrategy holds 499,096 Bitcoin worth $43.7 billion They've borrowed $8.2 billion to acquire it Their stock has crashed 55% in three months Their debt covenants have a "fundamental change" clause But here's what everyone misses: This isn't just about Bitcoin. It's about financial engineering that could collapse. But here's what everyone misses: This isn't just about Bitcoin. It's about financial engineering that could collapse. The Great Bitcoin Gamble Let's break down what's actually happening: MicroStrategy's entire business model relies on a flywheel effect: Borrow money through 0% convertible notes Buy Bitcoin and drive price higher Sell new shares at premium and buy more Bitcoin Repeat Borrow money through 0% convertible notes Buy Bitcoin and drive price higher Sell new shares at premium and buy more Bitcoin Repeat But here's where it gets interesting: This flywheel only works in one direction. But here's where it gets interesting: This flywheel only works in one direction. The Debt Structure Consider MicroStrategy's debt: $8.2 billion in convertible notes Most maturity dates between 2027-2032 $3 billion due in 2029 alone Conversion prices ranging from $39.80 to $672.40 $8.2 billion in convertible notes Most maturity dates between 2027-2032 $3 billion due in 2029 alone Conversion prices ranging from $39.80 to $672.40 These convertible notes contain a critical trigger – a "fundamental change" clause that could force early redemption if shareholders approve "any plan or proposal for the liquidation or dissolution of the Company." The Technical Trigger Here's the billion-dollar insight: If Bitcoin's value falls significantly, MicroStrategy's assets could slip below its liabilities, creating a perverse incentive for shareholders. If Bitcoin's value falls significantly, MicroStrategy's assets could slip below its liabilities, creating a perverse incentive for shareholders. Think about it: If Bitcoin falls 50%+ and stays down Convertible notes can't be converted (underwater) Debt comes due starting in 2027 Shareholders might prefer liquidation to pay themselves first If Bitcoin falls 50%+ and stays down Convertible notes can't be converted (underwater) Debt comes due starting in 2027 Shareholders might prefer liquidation to pay themselves first The Michael Saylor Factor Saylor currently holds 46.8% of voting power, making a liquidation vote difficult. He insists they wouldn't sell "even if Bitcoin fell to $1." But this ignores market realities: Creditors have legal rights Stock can't raise capital if it collapses The flywheel needs constant new money Cash flow can't cover interest payments Creditors have legal rights Stock can't raise capital if it collapses The flywheel needs constant new money Cash flow can't cover interest payments The Market Verdict MicroStrategy's stock has experienced wild swings: Up 500% in the Bitcoin bull run Down 55% in the recent correction Now trading around $252, down from $560+ in November Up 500% in the Bitcoin bull run Down 55% in the recent correction Now trading around $252, down from $560+ in November While the company's average Bitcoin purchase price is $66,357, their ability to service debt doesn't depend on being profitable – it depends on continued capital market access. The Real Questions The true vulnerabilities aren't about forced liquidation today. They're about: Can MicroStrategy refinance debt in a Bitcoin bear market? Will investors keep buying shares if Bitcoin falls further? What happens when convertible notes mature underwater? How sustainable is a "never sell" strategy with massive debt? Can MicroStrategy refinance debt in a Bitcoin bear market? Will investors keep buying shares if Bitcoin falls further? What happens when convertible notes mature underwater? How sustainable is a "never sell" strategy with massive debt? The Bitcoin Cult Factor As we noted in our analysis of Saylor's approach, he's positioned Bitcoin not as currency but as "digital gold" – a philosophical stance that's created cult-like devotion. The investment thesis relies on a permanent, upward trajectory. "If Bitcoin reaches $3M it would mean that something really bad is going on out there." But the flip side is troubling: What if it falls significantly and stays down? The Potential Endgame Three scenarios emerge: Bitcoin rises enough to make all convertible notes profitable, solving the problem Bitcoin trades sideways, creating refinancing challenges when debt matures Bitcoin falls significantly, potentially triggering a liquidation decision Bitcoin rises enough to make all convertible notes profitable, solving the problem Bitcoin rises enough Bitcoin trades sideways , creating refinancing challenges when debt matures Bitcoin trades sideways Bitcoin falls significantly , potentially triggering a liquidation decision Bitcoin falls significantly The Ultimate Irony Here's the greatest paradox: While Saylor positions Bitcoin as a hedge against monetary instability, he's built perhaps the most leveraged, fragile Bitcoin position possible. Welcome to the great MicroStrategy experiment. The question isn't whether Bitcoin has long-term value. The question is: Can a public company hold billions of it through debt without ever selling? The flywheel that took MicroStrategy to the moon works in both directions. And gravity is a powerful force.

Michael Saylor Has Built A House of Cards That Could Tumble Down

مايكل سيلور بنى بيتًا من ورق قابل للانهيار

মাইকেল সায়লর এমন এক তাসের ঘর তৈরি করেছেন যা ভেঙে পড়তে পারে

Michael Saylor ha construït un castell de cartes que podria caure

Michael Saylor ha construido un castillo de naipes que podría derrumbarse

マイケル・セイラーは崩れ落ちる可能性のあるトランプの家を建てた

Michael Saylor បានសាងសង់ផ្ទះកាតដែលអាចដួលរលំបាន។

Nanangana tranon'ny karatra mety hirodana i Michael Saylor

Michael Saylor Yubatse Inzu yamakarita Yashobora Kugwa

Michael Saylor postavil domček z kariet, ktorý by sa mohol zrútiť

Michael Saylor waxa uu dhisay guri kaadhadh ah oo dumi kara

Мајкл Сејлор направио је кућу од карата која би се могла срушити

Michael Saylor, Yıkılabilecek Bir Kart Evi İnşa Etti

Expertise Delivered Frontend, Backend & QA Industry Insurance Client Brief Our client, HealthPro Insurance, is a premier Medicare insurance broker, ranked among the top 5 in the United States by Healthcare.gov. Their comprehensive offerings include specialized support for health insurance, Dental, Vision, and long-term life insurance, ensuring their clients receive the best possible care and coverage. HealthPro provides valuable subsidies to employers, individual consumers, and nonprofits. They aim to enhance the accessibility and provision of insurance for all US citizens. Project Scope Our client managed their website and peripheral services on legacy processes and infrastructure. Their unoptimized processes and abrupt resource investments made them pay through their nose. They wanted to introduce automation with DevOps practices to optimize their system’s performance, flexibility, efficiency, and data security. HealthPro’s project scope revolves around the following attributes. Distribute server load and processes across multiple domains.
Switch to private cloud from public to prevent cyber attacks like DDoS, resulting in data breaches.
Creating two different environments for production and staging can mitigate the risks of compromised user experiences when experimenting with or upgrading systems. The Challenge Due to hosting all services in a single public cloud and relying solely on credential protection, our client faced server overloading and security vulnerabilities. This also increased the platform’s instability, posing limitations with hosting the production and staging environments. The above issues compromised the performance of their services, increased costs, and created data risk, hindering their accessibility and scalability. Another prime challenge was switching all their domains and sub-domains to AWS and implementing autoscaling to optimize costs. Our client’s services used MariaDP as a database, and a significant challenge was to plan a systematic switch from MariaDP to MySQL while keeping their data intact. Why Maruti Techlabs? One of our consultants, who was acquainted with our client's challenges and with our expertise, advocated for our services. Subsequently, the consultant and the client's Chief of Sales, who served as the primary point of contact, designed a questionnaire to evaluate our approach. After thoroughly understanding the client's expectations through the questionnaire, our team of experts brainstormed various solutions. We developed a detailed workflow outlining the implementation of these solutions, focusing on enhancing security, improving stability, and optimizing costs. Our proven DevOps expertise in server administration and maintenance, implementing and maintaining CI/CD pipelines, cloud migration, integrating DevSecOps, and more, supported by the consultant's confidence in our capabilities, secured the project for us. The proposed workflow and solutions demonstrated our ability to meet the client's needs effectively, ultimately earning their trust and business. Solution A phase-wise approach was chosen to implement the required solutions to ensure uninterrupted services for the client's customers. Here's how we implemented the perfect tech infrastructure and optimized costs for our client: 1. Isolating Staging and Production Environment Firstly, we isolated the whole(one) unit into two different environments for ‘Production’ and ‘Staging.’ This helped us offer a seamless customer experience in production, allowing our client to experiment with new features, load and regression testing, and introducing new upgrades with security in staging. 2. Transitioning from Public to Private Cloud Our next step was building the staging environment by creating a new Virtual Private Cloud (VPC) with private and public subnets. 3. Creating Data Back-Up We then transferred the database from the old VPC’s public subnet to the new VPC’s private subnet. We did this by taking a snapshot or backing up the data, creating a copy in the new VPC, and restoring it to the private subnet. Additionally, to allow the server to access the database seamlessly, we assigned roles to the EC2 server and upgraded its database security rules. 4. Cost Optimization As a final step, we optimized costs for our client by implementing automatic scaling with AWS’s Aurora Serverless. Aurora Serverless offers the convenience of auto-scaling CPU and memory usage based on traffic. For instance, if our clients don’t have weekly traffic, they do not need to pay for that period. We achieved this by migrating from MariaDB to MySQL, as Aurora supports both database engines. We then transferred the migrated data to Aurora Serverless. Communication and Collaboration We conducted a one-week discovery phase during which our team members communicated extensively to understand their current solution's limitations and expectations concerning their desired solution. We allocated two team members, one DevOps manager and one project manager, to execute this project’s specifics end-to-end. For the entirety of the project, our primary point of contact was our client’s ‘Chief of Sales.’ Here are the platforms we majorly use for communication: Slack - Quick communication,
Google Meet - Daily updates and queries, roll-outs
Email - Official channel to share contracts, weekly updates, and more. Technology Stack Result Implementing our DevOps solutions significantly improved the client's infrastructure and operational efficiency. Reduced on-platform search time from 8 minutes to under 1 minute.
We achieved a 300% boost in application performance over previous versions.
Our solution facilitated a 50% decrease in database and server management costs.
Optimized costs by implementing auto-scaling with AWS Aurora Serverless.
The device system performs automated backups and recovery in case of a crash or security threat.
In case of a sudden server load, the system sends alert signals through email and Slack.
Migration to a private cloud offers end-to-end data encryption. Our Development Process We follow Agile, Lean, & DevOps best practices to create a superior prototype that brings your users’ ideas to fruition through collaboration & rapid execution. Our top priority is quick reaction time & accessibility. We really want to be your extended team, so apart from the regular meetings, you can be sure that each of our team members is one phone call, email, or message away. Expertise Delivered Frontend, Backend & QA Industry Insurance Client Brief Our client, HealthPro Insurance, is a premier Medicare insurance broker, ranked among the top 5 in the United States by Healthcare.gov. Their comprehensive offerings include specialized support for health insurance, Dental, Vision, and long-term life insurance, ensuring their clients receive the best possible care and coverage. HealthPro provides valuable subsidies to employers, individual consumers, and nonprofits. They aim to enhance the accessibility and provision of insurance for all US citizens. Project Scope Our client managed their website and peripheral services on legacy processes and infrastructure. Their unoptimized processes and abrupt resource investments made them pay through their nose. They wanted to introduce automation with DevOps practices to optimize their system’s performance, flexibility, efficiency, and data security. HealthPro’s project scope revolves around the following attributes. Distribute server load and processes across multiple domains. Switch to private cloud from public to prevent cyber attacks like DDoS, resulting in data breaches. Creating two different environments for production and staging can mitigate the risks of compromised user experiences when experimenting with or upgrading systems. Distribute server load and processes across multiple domains. Switch to private cloud from public to prevent cyber attacks like DDoS, resulting in data breaches. Creating two different environments for production and staging can mitigate the risks of compromised user experiences when experimenting with or upgrading systems. The Challenge Due to hosting all services in a single public cloud and relying solely on credential protection, our client faced server overloading and security vulnerabilities. This also increased the platform’s instability, posing limitations with hosting the production and staging environments. The above issues compromised the performance of their services, increased costs, and created data risk, hindering their accessibility and scalability. Another prime challenge was switching all their domains and sub-domains to AWS and implementing autoscaling to optimize costs. Our client’s services used MariaDP as a database, and a significant challenge was to plan a systematic switch from MariaDP to MySQL while keeping their data intact. Why Maruti Techlabs? One of our consultants, who was acquainted with our client's challenges and with our expertise, advocated for our services. Subsequently, the consultant and the client's Chief of Sales, who served as the primary point of contact, designed a questionnaire to evaluate our approach. After thoroughly understanding the client's expectations through the questionnaire, our team of experts brainstormed various solutions. We developed a detailed workflow outlining the implementation of these solutions, focusing on enhancing security, improving stability, and optimizing costs. Our proven DevOps expertise in server administration and maintenance, implementing and maintaining CI/CD pipelines, cloud migration, integrating DevSecOps, and more, supported by the consultant's confidence in our capabilities, secured the project for us. The proposed workflow and solutions demonstrated our ability to meet the client's needs effectively, ultimately earning their trust and business. Solution A phase-wise approach was chosen to implement the required solutions to ensure uninterrupted services for the client's customers. Here's how we implemented the perfect tech infrastructure and optimized costs for our client: 1. Isolating Staging and Production Environment 1. Isolating Staging and Production Environment Firstly, we isolated the whole(one) unit into two different environments for ‘Production’ and ‘Staging.’ This helped us offer a seamless customer experience in production, allowing our client to experiment with new features, load and regression testing, and introducing new upgrades with security in staging. 2. Transitioning from Public to Private Cloud 2. Transitioning from Public to Private Cloud Our next step was building the staging environment by creating a new Virtual Private Cloud (VPC) with private and public subnets. 3. Creating Data Back-Up 3. Creating Data Back-Up We then transferred the database from the old VPC’s public subnet to the new VPC’s private subnet. We did this by taking a snapshot or backing up the data, creating a copy in the new VPC, and restoring it to the private subnet. Additionally, to allow the server to access the database seamlessly, we assigned roles to the EC2 server and upgraded its database security rules. 4. Cost Optimization 4. Cost Optimization As a final step, we optimized costs for our client by implementing automatic scaling with AWS’s Aurora Serverless. Aurora Serverless offers the convenience of auto-scaling CPU and memory usage based on traffic. For instance, if our clients don’t have weekly traffic, they do not need to pay for that period. We achieved this by migrating from MariaDB to MySQL, as Aurora supports both database engines. We then transferred the migrated data to Aurora Serverless. Communication and Collaboration We conducted a one-week discovery phase during which our team members communicated extensively to understand their current solution's limitations and expectations concerning their desired solution. We allocated two team members, one DevOps manager and one project manager, to execute this project’s specifics end-to-end. For the entirety of the project, our primary point of contact was our client’s ‘Chief of Sales.’ Here are the platforms we majorly use for communication: Slack - Quick communication, Google Meet - Daily updates and queries, roll-outs Email - Official channel to share contracts, weekly updates, and more. Slack - Quick communication, Google Meet - Daily updates and queries, roll-outs Email - Official channel to share contracts, weekly updates, and more. Technology Stack Result Implementing our DevOps solutions significantly improved the client's infrastructure and operational efficiency. Reduced on-platform search time from 8 minutes to under 1 minute. We achieved a 300% boost in application performance over previous versions. Our solution facilitated a 50% decrease in database and server management costs. Optimized costs by implementing auto-scaling with AWS Aurora Serverless. The device system performs automated backups and recovery in case of a crash or security threat. In case of a sudden server load, the system sends alert signals through email and Slack. Migration to a private cloud offers end-to-end data encryption. Reduced on-platform search time from 8 minutes to under 1 minute. We achieved a 300% boost in application performance over previous versions. Our solution facilitated a 50% decrease in database and server management costs. Optimized costs by implementing auto-scaling with AWS Aurora Serverless. The device system performs automated backups and recovery in case of a crash or security threat. In case of a sudden server load, the system sends alert signals through email and Slack. Migration to a private cloud offers end-to-end data encryption. Our Development Process We follow Agile, Lean, & DevOps best practices to create a superior prototype that brings your users’ ideas to fruition through collaboration & rapid execution. Our top priority is quick reaction time & accessibility. We really want to be your extended team, so apart from the regular meetings, you can be sure that each of our team members is one phone call, email, or message away.

Reducing Server and Database Costs by 50% for an Insurance Broker using AWS

تخفيض تكاليف الخادم وقاعدة البيانات بنسبة 50% لشركة وساطة تأمين تستخدم AWS

Servidor ukat Base de Datos ukan qullqip 50% ukj jisk’achaña mä Corredor de Seguros ukan AWS apnaqasa

AWS-dən istifadə edən Sığorta Brokeri üçün Server və Verilənlər Bazası Xərclərinin 50% azaldılması

AWS ব্যবহারকারী বীমা ব্রোকারের জন্য সার্ভার এবং ডাটাবেস খরচ ৫০% কমানো

Reducción de costos de servidor y base de datos en un 50% para un corredor de seguros que utiliza AWS

Palvelin- ja tietokantakustannusten alentaminen 50 % AWS:ää käyttävän vakuutusmeklarin osalta

AWS を利用して保険ブローカーのサーバーとデータベースのコストを 50% 削減

Kokitisa ba Coûts ya Serveur na Base de Données na 50% pona Correcteur d'Assurance oyo azali kosalela AWS

Obniżenie kosztów serwera i bazy danych o 50% dla brokera ubezpieczeniowego korzystającego z AWS

Kugabanya Seriveri na Ububikoshingiro Ibiciro 50% kuri Broker wubwishingizi ukoresheje AWS

Minska server- och databaskostnader med 50 % för en försäkringsmäklare som använder AWS

ንሓደ AWS ዝጥቀም ናይ መድሕን ደላላይ ዋጋ ሰርቨርን ዳታቤዝን ብ50% ምንካይ

Table of Links Abstract and 1. Introduction 2 Related Work 2.1 The Impact of Blockchain Technology on Education 2.2 Incorporating Scratch into Tinker Learning for Computational Thinking 2.3 Using Polygon SDK for JavaScript to Interact with a Smart Contract 2.4 Challenges in Blockchain Education and the Need for User-friendly Tools 3 Methods 4 Results 5 Discussion 6 Conclusions and References 3 METHODS Our paper proposes a novel method for updating the metadata of non-fungible tokens (NFTs) using an API layer to reduce barriers to entry and enable K-12 students to explore blockchain technology. Our approach leverages Google Blockly, a visual programming language, to make updating NFT metadata more accessible and engaging for young learners. By creating a custom block in Blockly using the Blockly Developer Tools, we can create an interface that generates JavaScript and XML code for the block. This interface should allow the block’s JavaScript code to interact directly with the Polygon software development kit (SDK) and smart contracts. By integrating this custom block into the BlocklyDuino editor, K-12 learners can manipulate data collected from Arduino sensors and upload it to the Polygon blockchain using the custom block. In our paper, we present a method for creating custom blocks in Google Blockly and using them to update NFT metadata. First, we create a custom block in Google Blockly that outputs the "Hello World" message. Then, we name the block and add a dummy input before visually representing the custom block and adding it to the MyBlocks category. Next, we write a code generator function that generates the code required to update NFT metadata when the custom block is used. We modify the code to update the metadata with the desired information, export the custom block toolbox and import it into BlocklyDuino. This open-source visual programming platform generates Arduino code, and we also provide all the necessary files required for running the custom block. Finally, we import the toolbox into BlocklyDuino and verify that the custom block can be used to update NFT metadata. Our proposed method could promote student engagement and understanding of blockchain technology and NFTs, which could have significant implications forthe future of education. By leveraging a familiar and engaging visual programming language like Google Blockly, we hope to inspire a new generation of blockchain enthusiasts and foster a deeper understanding of digital ownership and value. Our paper’s third authoris a seventh-grade student responsible for following the teaching steps we provided to interact with Blockly and smart contracts to create an IoT plant watering and growing log. He used an NFT as a digital identity for the observed plant and regularly uploaded data to the blockchain while updating the NFT’s attributes. This demonstrates that our proposed method significantly lowers the barrier to entry for K-12 students to enter the blockchain world. Currently, our focus is on validating the process, but we plan to design more new course materials that create value for blockchain applications in the future. The Google Blockly Custom Block Tutorial is from creating the "Hello World" block and using it in BlocklyDuino. This tutorial will guide us through creating our first custom block in Google Blockly and integrating it into BlocklyDuino. The "Hello World" block will output "Hello World" in BlocklyDuino. Google Blockly demo (Custom block development tool): https://blocklydemo.appspot.com/static/demos/blockfactory/index.html. Creating the First Custom Block in Google Blockly "Hello World." The approach for creating custom blocks in Blockly is as follows: (1) Custom block in the "Blockly Developer Tools" and name the custom block "helloworld." https://blockly-demo.appspot. com/static/demos/blockfactory/index.html. (2) Add a dummy input and use the Block Factory to design our custom block. We can set the block’s type, input, output, and other properties. (3) Generate code. We can generate the code for it using the Block Factory. Teachers can help students add the Polygon SDK in the step. (4) Import the custom block’s XML definition into our BlocklyDuino workspace. https://qiudaru.github.io/webduino/. (5) Use the custom block in our BlocklyDuino programs by dragging it from the toolbox and connecting it to other blocks. Following these steps, we can create custom blocks in Blockly to add new functionality to the visual programming language. Google Blockly is a visual programming language that allows users to create JavaScript code by selecting and connecting blocks representing desired functionality [? ]. This involves choosing the desired block, dragging and dropping it into the workspace, linking the blocks, configuring them by setting properties such as text, color, and tooltip, generating the JavaScript code, and exporting it to a JavaScript file or copying and pasting it into a project. Users can then test and refine the code to improve its functionality and performance. By pushing Blockly JavaScript files to a GitHub repository and configuring BlocklyDuino to access the repository’s page, users can create programs that set data to a smart contract with Polygon SDK [? ]. This involves creating a GitHub repository to store the Blockly JavaScript files, pushing the files to the repository, As shown in figure 2, the Block Definition on the right side defines the block. As shown in figure 2, the generated code in the "Generator stub" is: creating a GitHub page for the repository, configuring BlocklyDuino to access the GitHub page, creating a program in Blockly that sets data to a smart contract with Polygon SDK, generating the JavaScript code, uploading the code to an Arduino board using BlocklyDuino, connecting the board to the smart contract, and executing the program to set the data. Authors:
(1) Yun-Cheng Tsai, National Taiwan Normal University, Taipei, Taiwan (pecu@ntnu.edu.tw);
(2) Jiun-Yu Huang, Small Town Intelligence Co., Ltd, Taiwan (yillkid@gmail.com);
(3) Da-Ru Chiou, Taitung Tung Hai Junior High School, Taiwan (1234567si2k@gmail.com). This paper is available on arxiv under CC BY 4.0 DEED license. Table of Links Abstract and 1. Introduction Abstract and 1. Introduction 2 Related Work 2.1 The Impact of Blockchain Technology on Education 2.1 The Impact of Blockchain Technology on Education 2.2 Incorporating Scratch into Tinker Learning for Computational Thinking 2.2 Incorporating Scratch into Tinker Learning for Computational Thinking 2.3 Using Polygon SDK for JavaScript to Interact with a Smart Contract 2.3 Using Polygon SDK for JavaScript to Interact with a Smart Contract 2.4 Challenges in Blockchain Education and the Need for User-friendly Tools 2.4 Challenges in Blockchain Education and the Need for User-friendly Tools 3 Methods 3 Methods 4 Results 4 Results 5 Discussion 5 Discussion 6 Conclusions and References 6 Conclusions and References 3 METHODS Our paper proposes a novel method for updating the metadata of non-fungible tokens (NFTs) using an API layer to reduce barriers to entry and enable K-12 students to explore blockchain technology. Our approach leverages Google Blockly, a visual programming language, to make updating NFT metadata more accessible and engaging for young learners. By creating a custom block in Blockly using the Blockly Developer Tools, we can create an interface that generates JavaScript and XML code for the block. This interface should allow the block’s JavaScript code to interact directly with the Polygon software development kit (SDK) and smart contracts. By integrating this custom block into the BlocklyDuino editor, K-12 learners can manipulate data collected from Arduino sensors and upload it to the Polygon blockchain using the custom block. In our paper, we present a method for creating custom blocks in Google Blockly and using them to update NFT metadata. First, we create a custom block in Google Blockly that outputs the "Hello World" message. Then, we name the block and add a dummy input before visually representing the custom block and adding it to the MyBlocks category. Next, we write a code generator function that generates the code required to update NFT metadata when the custom block is used. We modify the code to update the metadata with the desired information, export the custom block toolbox and import it into BlocklyDuino. This open-source visual programming platform generates Arduino code, and we also provide all the necessary files required for running the custom block. Finally, we import the toolbox into BlocklyDuino and verify that the custom block can be used to update NFT metadata. Our proposed method could promote student engagement and understanding of blockchain technology and NFTs, which could have significant implications forthe future of education. By leveraging a familiar and engaging visual programming language like Google Blockly, we hope to inspire a new generation of blockchain enthusiasts and foster a deeper understanding of digital ownership and value. Our paper’s third authoris a seventh-grade student responsible for following the teaching steps we provided to interact with Blockly and smart contracts to create an IoT plant watering and growing log. He used an NFT as a digital identity for the observed plant and regularly uploaded data to the blockchain while updating the NFT’s attributes. This demonstrates that our proposed method significantly lowers the barrier to entry for K-12 students to enter the blockchain world. Currently, our focus is on validating the process, but we plan to design more new course materials that create value for blockchain applications in the future. The Google Blockly Custom Block Tutorial is from creating the "Hello World" block and using it in BlocklyDuino. This tutorial will guide us through creating our first custom block in Google Blockly and integrating it into BlocklyDuino. The "Hello World" block will output "Hello World" in BlocklyDuino. Google Blockly demo (Custom block development tool): https://blocklydemo.appspot.com/static/demos/blockfactory/index.html. Creating the First Custom Block in Google Blockly "Hello World." The approach for creating custom blocks in Blockly is as follows: (1) Custom block in the "Blockly Developer Tools" and name the custom block "helloworld." https://blockly-demo.appspot. com/static/demos/blockfactory/index.html. (2) Add a dummy input and use the Block Factory to design our custom block. We can set the block’s type, input, output, and other properties. (3) Generate code. We can generate the code for it using the Block Factory. Teachers can help students add the Polygon SDK in the step. (4) Import the custom block’s XML definition into our BlocklyDuino workspace. https://qiudaru.github.io/webduino/. (5) Use the custom block in our BlocklyDuino programs by dragging it from the toolbox and connecting it to other blocks. Following these steps, we can create custom blocks in Blockly to add new functionality to the visual programming language. Google Blockly is a visual programming language that allows users to create JavaScript code by selecting and connecting blocks representing desired functionality [? ]. This involves choosing the desired block, dragging and dropping it into the workspace, linking the blocks, configuring them by setting properties such as text, color, and tooltip, generating the JavaScript code, and exporting it to a JavaScript file or copying and pasting it into a project. Users can then test and refine the code to improve its functionality and performance. By pushing Blockly JavaScript files to a GitHub repository and configuring BlocklyDuino to access the repository’s page, users can create programs that set data to a smart contract with Polygon SDK [? ]. This involves creating a GitHub repository to store the Blockly JavaScript files, pushing the files to the repository, As shown in figure 2, the Block Definition on the right side defines the block. As shown in figure 2, the generated code in the "Generator stub" is: creating a GitHub page for the repository, configuring BlocklyDuino to access the GitHub page, creating a program in Blockly that sets data to a smart contract with Polygon SDK, generating the JavaScript code, uploading the code to an Arduino board using BlocklyDuino, connecting the board to the smart contract, and executing the program to set the data. Authors: (1) Yun-Cheng Tsai, National Taiwan Normal University, Taipei, Taiwan (pecu@ntnu.edu.tw); (2) Jiun-Yu Huang, Small Town Intelligence Co., Ltd, Taiwan (yillkid@gmail.com); (3) Da-Ru Chiou, Taitung Tung Hai Junior High School, Taiwan (1234567si2k@gmail.com). Authors: Authors: (1) Yun-Cheng Tsai, National Taiwan Normal University, Taipei, Taiwan (pecu@ntnu.edu.tw); (2) Jiun-Yu Huang, Small Town Intelligence Co., Ltd, Taiwan (yillkid@gmail.com); (3) Da-Ru Chiou, Taitung Tung Hai Junior High School, Taiwan (1234567si2k@gmail.com). This paper is available on arxiv under CC BY 4.0 DEED license. This paper is available on arxiv under CC BY 4.0 DEED license. available on arxiv available on arxiv

Stories, News & Mentions about PLANA

Plana’s Stories on HackerNoon

Plana's News & Mentions