Amitesh Srivastava

Thank you, Robert and Justin, for insightful discussion. In the 𝐌𝐞𝐚𝐧𝐂𝐚𝐜𝐡𝐞 study, we argue that traditional cache hit and miss rates are misleading for 𝒔𝒆𝒎𝒂𝒏𝒕𝒊𝒄 𝒄𝒂𝒄𝒉𝒊𝒏𝒈 and should not be used. MeanCache is a user-centric semantic cache for LLM-based web services (e.g., ChatGPT). It uses Federated Learning to fine-tune smaller embedding models, enhancing query semantic matching and handling context queries effectively. 𝐏𝐚𝐩𝐞𝐫: https://lnkd.in/e9PB6qie

13

Over the last 1 year, I've dived deep into the codebases of major projects like TensorFlow, MongoDB, Chromium, OpenSSL, LLVM, GCC, Kafka, Qt, CMake, and currently the Linux kernel. Transitioning from data structures and Development using Javascript based Tech Stacks to exploring open source C++ projects felt like stepping into an entirely new world. What I discovered were common practices and patterns that define production-level C++ code across these projects. Curious to know what they are? I've compiled my insights into a PDF—perfect for anyone looking to explore the Open Source Projects and contribute to the same. #C++ #OpenSource #SoftwareDevelopment #Programming

86

8 Comments

I'm reaching out to my LinkedIn community during a time of unexpected change. Recently, I found myself among the many affected by Tesla layoffs. After 16 months of pouring my passion and skills into projects I truly believed in, saying goodbye was not easy. During my tenure on the Vision Automation team, I had the privilege of: * Enhancing vehicle quality by using computer vision to identify defects. * Architecting an in-house machine learning platform for efficient model labelling, training, and deployment. * Developing a robust labeling platform to gather SME insights and train cutting-edge ML models. * Streamlining our authentication processes by transitioning to Microsoft OIDC. * Creating an application alert system by customizing Prometheus, Grafana, and Alert Manager for application-specific alarms. * Collaborating with an extraordinary team of engineers who inspired me every single day. Now, as I navigate this transition on a visa with a tight 60-day timeline, I am actively seeking opportunities where I can bring my expertise to the table. I'm looking for roles in Software Engineering, Distributed Systems, or Machine Learning Operations where I can continue to grow and make an impact. If you, or someone you know, can guide me towards potential opportunities, I would be grateful for your support. Let's connect and explore how my experience can benefit your team. Together, we can achieve great things! #TeslaLayoffs #OpenToWork #Tesla #EngineeringExcellence #MachineLearning #SoftwareEngineering

1,834

101 Comments

I couldn’t find any cheat sheet on system design tradeoffs, so I created one myself. So, a few weeks ago, I was learning more about system design tradeoffs in distributed systems. No resource gave me all the information in one place, so I decided to make one myself. Here are the links to my learnings: ►𝐀𝐜𝐜𝐮𝐫𝐚𝐜𝐲 𝐯𝐬. 𝐋𝐚𝐭𝐞𝐧𝐜𝐲: https://lnkd.in/gGyTb3Jh ►𝐓𝐡𝐫𝐨𝐮𝐠𝐡𝐩𝐮𝐭 𝐯𝐬. 𝐋𝐚𝐭𝐞𝐧𝐜𝐲: https://lnkd.in/gzNb85xY ►𝐏𝐮𝐥𝐥 𝐯𝐬. 𝐏𝐮𝐬𝐡 𝐀𝐫𝐜𝐡𝐢𝐭𝐞𝐜𝐭𝐮𝐫𝐞: https://lnkd.in/gKaaG-d6 ►𝐌𝐞𝐦𝐨𝐫𝐲 𝐯𝐬 𝐋𝐚𝐭𝐞𝐧𝐜𝐲: https://lnkd.in/gGJ_SSDT ►𝐒𝐐𝐋 𝐯𝐬 𝐍𝐨𝐒𝐐𝐋 𝐝𝐚𝐭𝐚𝐛𝐚𝐬𝐞𝐬: https://lnkd.in/g6QCyZga – P.S: I write a lot about distributed systems and system design; let me know if there’s anything you’d like me to cover…

244

5 Comments

Recently one of the stories regarding investing in Intel Corporation is going viral. So apparently a college math major got a $800k inheritance from his grandmom and he used $700k of that to invest in INTC hours before the Q2 2024 earnings were out. I know, a terrible choice in hindsight! 🙃 And he lost $200k of that within hours after the terrible report came out and management expressed cluelessness. That too on stocks, he was not even doing derivative leveraged play like options. 😅 This story is a very good lesson that stocks inherently are a risky investment, and it is good to be diversified to some extent. Some lessons to keep in mind: 1. Do not put all your eggs in one basket unless you know the business inside out. 2. Always do your due diligence on the company and the management. Company being well known is not enough, it is also about the ability of the management to execute. 3. Just buying low P/E stocks is not value investing and is not safe by any means. #investing #equity #stocks

26

4 Comments

Excited to share my latest tech blog post on expansively training #robotics on #AWS using #LLM! Big thank you to my amazing team members for their support and insights. Check it out for insights into 1. how to leverage #LLMs on Amazon #Bedrock to craft reward functions; 2. how to scale robotic #simulation and training on Amazon #EKS; and 3. how to visualize simulation applications running in Amazon #EKS. Would love to hear your thoughts!

14

10x engineer, rock star, brogrammers....these are all unhelpful myths. Were they a great dev? Probably, but a great dev should not have to suffer through ego death to consider themselves successful. I agree with others, this attitude is not how a team succeeds and it is toxic. When I look at Facebook now I see an application stack that constantly breaks, is unintuitive and doesn't have the infrastructure to support itself, worse in quality than it was a decade ago. Exactly the type of thing I would expect from a bunch of individual contributors without cohesion. I'm sure this analysis on my part is unfair and there are dedicated, hard working team members there ready to make a difference, but this type of culture needs to stop. Invest in your internal teams, promote from within, train, encourage, build great things together instead of waiting for an imaginary deus ex programmator to save you.

7

1 Comment

Just stumbled upon this article. For those who are in sophomore year of their graduation and interested in ML and Gen AI. #AI #ML #GENAI

9

The Power of AI-Driven Pair Programming As a Staff Engineer at Artium, I've witnessed firsthand the transformative power of integrating AI tools, particularly Large Language Models (LLMs), into pair programming. This innovative approach is revolutionizing the way we code and collaborate, and I'm excited to share my insights with you. One of the most significant benefits of utilizing LLMs like GPT in pair programming is the dramatic acceleration of coding processes. By providing real-time feedback and code suggestions, these AI models reduce the time spent on routine tasks and debugging, allowing teams to push updates and features at an unprecedented pace. This increased speed and efficiency is a game-changer in the fast-paced world of software development. Moreover, LLMs leverage vast datasets to suggest coding solutions that adhere to best practices, significantly improving the quality and reliability of software products. This AI-driven guidance helps minimize errors and enhance overall code performance, ensuring that we deliver top-notch solutions to our clients. Another crucial aspect of AI-driven pair programming is the optimal utilization of resources. By handling repetitive tasks, AI allows developers to focus on complex problem-solving and creative innovation. This shift not only optimizes human staff allocation but also amplifies the creative potential of development teams, leading to more groundbreaking solutions. AI models in pair programming also serve as an on-demand knowledge base, teaching new frameworks and languages, and keeping the team updated with the latest industry standards. This fosters a culture of continuous learning and professional growth, which is essential in the ever-evolving tech landscape. Furthermore, with AI, the benefits of pair programming are no longer confined to office hours. Developers can interact with AI partners anytime, ensuring continuous productivity and making software development more adaptable to varying work schedules. This flexibility is crucial in today's diverse and dynamic work environments. In my experience, the AI-human partnership in software development not only increases operational efficiency but also paves the way for more innovative solutions. As we continue to explore and understand AI's potential, it's clear that the fusion of AI with human ingenuity in pair programming is just the beginning of a larger transformative journey in tech. I'm curious to hear your thoughts and experiences. What AI tools have you integrated into your development process, and how have they transformed your workflow? Let's discuss how we can further harness AI to enhance our productivity and creativity in tech. #SoftwareDevelopment #PairProgramming #AIinTech #Innovation #TechCommunity

5

Fun fact: StubHub, a multi-billion dollar e-commerce platform, is run by fewer than 200 engineers. 👨‍💻 👩‍💻 Many people I meet at networking events are surprised by this, yet it truly highlights the concentration of talent at our company. At StubHub, each engineer juggles multiple roles every week. From data engineering to working with large language models, developing backend system APIs, crafting Snowflake SQL procedures for analytics, to enhancing software observability—all while being on a rotating on-call schedule. This versatility not only fosters exponential growth and effective time management but also teaches us how to efficiently segment large projects into manageable weekly or quarterly parts. #TalentDensity #TechInnovation #SoftwareEngineering #DataEngineering #AI #ProjectManagement

96

11 Comments

Don’t be fooled. There isn’t a “best” kind of SWE. Only the right one for the job. Product devs can drive insane impact on greenfield platforms. - Instagram gained its first 14 million users with only 3 engineers. And infra gurus can save gobs of money optimizing performance at scale. - 1 Mixpanel dev cut their Google BigQuery spend by 80% in a matter of weeks. If you’re struggling to stand out, ask yourself: “What kind of engineer does my team need 6 months from now?” Then become that person.

213

44 Comments

Here is my interview experience for the Software Engineer 2 role at Nordstrom, where I went through a series of diverse and challenging round. 💻 Round 1 - Coding round: The first round focused on linked list operations, with a question about merging two sorted linked lists. I had to implement an optimized solution that combined the lists while maintaining their order, ensuring efficient time and space complexity. This round tested my ability to manage pointer manipulation and handle traversal techniques effectively. 💻 Round 2 - Coding round: The second round covered graph traversal using Depth First Search and Breadth First Search algorithms. I solved problems involving both recursive and iterative approaches to graph traversal. It emphasized my ability to explore different strategies for complex graph structures. 🛠️ Round 3 - System Design : I was asked to design a real-time commenting system similar to Facebook Live. We discussed scaling for large volumes of traffic, ensuring low-latency communication, and managing database consistency. This round focused on building a robust architecture with optimal trade-offs. 🤝 Round 4 - Cross-Functional Team Interview: This round involved collaboration with product management and design, with a focus on questions related to teamwork and communication. I was asked about my experience collaborating with engineering and non-engineering teams, along with situation-based questions on resolving conflicts and aligning team goals. This round highlighted my ability to work effectively across diverse teams and manage real-world challenges. 🤝 Round 5 - Behavioral Interview: This round was centered on discussing my work experiences and handling challenges in a team setting. I shared examples in the STAR format, emphasizing my adaptability, problem-solving, and leadership skills. It provided a platform to demonstrate alignment with Nordstrom’s values and culture. Each round of the interview process helped sharpen both my technical and interpersonal skills. I'm grateful for the opportunity to interview with Nordstrom and for the valuable insights I gained along the way! #TechInterviews #NordStrom #SoftwareEngineer #CodingChallenges #InterviewExperience #TechJobs #SystemDesign

111

2 Comments

During my short time with NVIDIA's Display Drivers team, I had the opportunity to build a chatbot that helped developers quickly access information related to sanity infrastructure. I also attended demos, like NVIDIA's ChatRTX(https://lnkd.in/gsU4x7Vt), that were building similar RAG chatbots. Recently, I found a course, that covered the challenges I faced such as pre-processing diverse data sources and retaining memory context, in just an hour(or a weekend if you code along). Things are moving quite fast for LLMs. While looking for resources to further bridge my knowledge gaps and work on another project, I came across another gem from DeepLearning.AI. For those interested in learning about LLMs, I highly recommend(in the same order): 1. Generative AI with LLMs(https://lnkd.in/guQz4Mcc): Key LLM concepts and hands-on experience via a Dialogue Summarization project. 2. LangChain Chat with Your Data(https://lnkd.in/gRUQVdSR): A practical guide to building chatbots that interact with your data, covering common challenges and solutions. #llm #machinelearning #deeplearning #AI #artificialintelligence #generativeai

36

Here’s a quick summary of how a Trump administration could affect high-skilled immigration in the U.S. 1. 𝐌𝐨𝐫𝐞 𝐇-1𝐁 𝐕𝐢𝐬𝐚 𝐃𝐞𝐧𝐢𝐚𝐥𝐬: The denial rate for H-1B work visas (for skilled workers) might go up. Under Trump, it was around 18%, compared to just 3% recently. 2. 𝐌𝐨𝐫𝐞 𝐄𝐱𝐭𝐫𝐚 𝐏𝐚𝐩𝐞𝐫𝐰𝐨𝐫𝐤 𝐟𝐨𝐫 𝐇-1𝐁𝐬: There could be more "requests for evidence" (RFEs) where applicants need to provide extra documents. Under Trump, the rate was higher (34%) than it is now (12%). 3. 𝐅𝐞𝐰𝐞𝐫 𝐖𝐨𝐫𝐤 𝐏𝐞𝐫𝐦𝐢𝐭𝐬 𝐟𝐨𝐫 𝐇-1𝐁 𝐒𝐩𝐨𝐮𝐬𝐞𝐬: Trump had tried to stop spouses of H-1B workers from being able to work. This might happen again. 4. 𝐋𝐨𝐰𝐞𝐫 𝐀𝐩𝐩𝐫𝐨𝐯𝐚𝐥 𝐟𝐨𝐫 𝐒𝐨𝐦𝐞 𝐎𝐭𝐡𝐞𝐫 𝐖𝐨𝐫𝐤 𝐕𝐢𝐬𝐚𝐬: Certain visa types like EB-1 (for extraordinary skills) and L-1 (for company transfers) might see fewer approvals, based on Trump’s previous approach. Other categories, like EB-2, didn’t change much last time. 5. 𝐒𝐭𝐮𝐝𝐞𝐧𝐭 𝐖𝐨𝐫𝐤 𝐏𝐫𝐨𝐠𝐫𝐚𝐦𝐬 𝐂𝐨𝐮𝐥𝐝 𝐛𝐞 𝐋𝐢𝐦𝐢𝐭𝐞𝐝: Trump previously aimed to reduce work opportunities for international students on F-1 visas after graduation. 6. 𝐇𝐢𝐠𝐡𝐞𝐫 𝐖𝐚𝐠𝐞 𝐑𝐞𝐪𝐮𝐢𝐫𝐞𝐦𝐞𝐧𝐭𝐬 𝐟𝐨𝐫 𝐇-1𝐁: Trump proposed raising minimum wages for H-1B holders to make it harder for lower-paid positions to qualify. Although Trump mentioned supporting green cards for U.S. college graduates, his policies on high-skilled immigration were generally stricter than recent policies.

27

7 Comments

Here’s how DoorDash evaluates its RAG performance by using LLM As A Judge. It's able to monitor retrieval correctness, response accuracy, coherence, and relevance to support questions which are posed by Dashers. 1. First, thousands of chat transcripts from Dashers are reviewed to identify common issues. 2. The LLM Judge is prompted with open-ended questions to assess them, and its answers are summarised into actionable feedback. 3. From these, high-frequency issues are built into prompts or rule-based metrics for continuous monitoring. 4. Human review teams validate the LLM’s evaluations by reviewing random transcript samples. This human feedback is calibrated with the automated system for better coverage. LLM Judge helps address issues like inaccurate retrieval, hallucinations, and low quality prompts. It makes sure that there's a continuous quality improvement pipeline alongside human support agents. Link to their full report is in the comments. #AI #LLMs #RAG

90

3 Comments

I recently learned about Finetuning Large Language Models on DeepLearning.AI x Lamini taught by Sharon Zhou, PhD. Key Takeaways: 1. Understand when to apply finetuning on LLMs. 2. Prepare your data for finetuning. 3. Train and evaluate an LLM on your data. With fine-tuning, you can take your data to train the model on it and update the weights of the neural nets in the LLM, changing the model compared to other methods like prompt engineering and Retrieval Augmented Generation. Finetuning allows the model to learn style, and form and can update the model with new knowledge to improve results. #finetuning #datascience #machinelearning #deeplearning #generativeai #llm #promptengineering #pytorch #llama #meta

31

1 Comment

🚀 Excited to Explore New Opportunities! 🚀 Over the past few years, I’ve had the privilege to work with leading companies like Amazon Web Services (AWS), Rivian, Intel, and OYO. Although some transitions were shorter than I initially expected—due to organizational changes and market conditions—each role allowed me to grow, make meaningful contributions, and sharpen my skills across backend, frontend, and full-stack development. These experiences have shaped my adaptability, resilience, and ability to hit the ground running. With expertise in Go, Java, React.js, AWS tools, microservices, and data-driven architectures, I am eager to find a long-term role where I can grow within the organization, contribute to impactful projects, and build a lasting partnership. 🔗 Key Skills & Technologies: Languages: Go, Python, JavaScript/Typescript, Java Databases: MongoDB, Redis, MySQL, Elasticsearch, and more. Technologies: React.js, Node.js, AWS, GraphQL, RPC, and more Interests: Full-Stack Engineering, Front-end/Backend Engineering, Machine Learning, Data Science I am actively seeking roles in Software Development, Backend Engineering, or Full-Stack Development in organizations with dynamic environments. If your network or organization is hiring for roles that align with my experience and aspirations, I’d love to connect. This time, I am particularly focused on finding a role where I can make a lasting impact and grow alongside the organization. Thank you for your support and referrals! Feel free to reach out via LinkedIn or my Portfolio page (The link is on my profile > featured section), or share this post with your network. #OpenToWork #SoftwareEngineering #Frontend #FrontendDeveloper #BackendDeveloper #FullStackDevelopment #AWS #GoLang #JavaScript #CareerGrowth #LongTermCommitment #Microservices #JobSearch #TechJobs #Networking

44

5 Comments