Vishwas Tripathi

🍀 6 months at Google.... Today I have completed 6 months at Google after having joined back in January 27th, 2025. These 6 months have been a rollercoaster of learning, exploration, meeting some exceptionally skilled people, and growth. Summarizing my journey at Google, here's the top 6 learnings for me over the period of the last 6 months... 1️⃣. Imposter Syndrome is real - Working at a place like Google, the talent density is the real deal. With being surrounded by some the smartest folks in the industry and working alongside them, its easy to feel lost and out of place. 2 things kept me going here - a. The support from the team in my rampup. b. The confidence that I've done harder things in the past - this too shall pass. 2️⃣. Ownership matters - At a place like Google, with highly skilled people all around, its kinda obvious that people like to own their work and expect you to do the same. How to own your work end to end, and to deliver it with the highest quality is something I've greatly learnt in the past few months. 3️⃣. Communication >>> Code - At a place, where any code change you make affects billions, communication and alignment regarding the same becomes something of utmost nessecity. Skillful communication within the team, cross teams, and with multiple stakeholders to effectively deliver work is something I've seen matter more than ever here. 4️⃣. Asking for help is a strength - In companies like Google and Microsoft, asking for help is looked on as a strength, not a weakness. The culture fosters that. However, to get help, you've to ask for it. Getting over the thought of judgment, to reach out for help, is a key learning which helped me a lot in the journey. 5️⃣. Make meaningful connections - Walking in any journey, what makes it worthwhile are the friends you make along the way. Making intentional connections, and going out of the way to help others move ahead in their journey, is the way to move ahead in the journey. This is something I also try to incorporate in my Linkedin journey by trying to give back to the community. 6️⃣. You're never too good at something, Keep learning - Tech is a fast moving space. The thought of "I already know enough" becomes obsolete as fast as the tech of today. Its a journey of constant of learning and upskilling. Parallely, make your basics stronger, that's what allows you to keep up the learning pace, with the industry. The last 6 months have been really exciting, fun, full of learnings and challenges all along the way. As its said in Google, being a Noogler is like "trying to drink from a firehose". Really excited for what the next 6 months unfolds. Stay tuned for more such content!!✌🏻🚀

908

85 Comments

India may have just stopped playing catch-up in AI. 🇮🇳 For years, the LLM conversation felt like a two-horse race: Silicon Valley 🇺🇸 , Beijing 🇨🇳 If you wanted a serious foundation model, you looked West or East. BUT, something interesting happened. Sarvam released 30B and 105B parameter models — and they weren’t built by wrapping someone else’s model. No quick fine-tunes. No “Llama with a new name”. These models were trained from scratch. Even more interesting: they support 22 Indian languages natively. That matters more than it sounds. Because AI built for India shouldn’t just translate our languages. It should understand the context, dialects, and cultural nuances behind them. For a long time, India has been one of the largest users of AI products. Now we’re starting to build the foundations too. And that could change the entire AI ecosystem coming out of this country. Exciting times for builders. HF Link: https://lnkd.in/gJ_Rk9kV (105b) https://lnkd.in/gp3W32RD (30b) #SarvamAI #IndiaTech #OpenSourceAI #GenerativeAI #MakeInIndia

26

9 Comments

🚨 SWE in 2025 and still ignoring LLMs? You are not just behind …. You are replaceable 💣 But the good news? It’s not too late to catch up — and you don’t need a fancy degree or paid course to start. Note: This is not promotional! I found a goldmine of a FREE resource by Victor Dantas that takes you from: 🔸 Basic Machine Learning ➡️ to Neural Networks ➡️ to RNNs ➡️ to Transformers ➡️ all the way up to LLMs like GPT & LaMDA. It’s literally a zero-to-hero roadmap that breaks down how these complex systems actually work — in plain English. 📚 Whether you’re a SWE, student, or AI-curious engineer — this is the kind of stuff that will future-proof your career. #AI #LLM #MachineLearning #NeuralNetworks #Transformers #SoftwareEngineering #GPT #LearnForFree

190

37 Comments

Competitive programming is a waste of time for you… if you are not polishing your development side by side. Unless you are targeting quant companies, grinding CP alone will not get you into the top product companies everyone talks about. Yes, it helps you clear DSA rounds. But your job as a software engineer will demand much more than solving problems on a timed platform. And in this market, that gap shows very fast. Interviews test problem-solving. Jobs test system thinking, debugging, ownership, communication, and shipping. If you are great at CP but cannot: • Build and deploy a backend • Design clean APIs • Work with databases and caching • Debug production issues • Read and improve someone else’s code • Explain your design decisions clearly You will struggle even after clearing initial rounds. Now look at the people landing back-to-back internships. Most of them have: • Decent DSA • Strong development fundamentals • Working projects they can demo • Proof that they can ship and maintain code • Some exposure to system design They may not top leaderboards. But they look dependable to hiring managers. And dependable engineers get hired. The biggest mistake students make is this: They treat CP as the destination. It was only meant to be a tool. CP sharpens how you think. Development decides how useful you become. If your entire identity is DSA, an off interview can damage your confidence. If your identity is building things, interviews become just one more step. So if you are grinding CP right now, do this in parallel: • Build one proper product • Deploy at least one full stack project • Learn how APIs fail • Learn how databases break • Learn what happens under load That is how problem-solving turns into employability. DSA helps you enter the door. Development decides how far you go.

434

21 Comments

While reading a paper, I came across the term multi-socket architecture - an extreme case of vertical scaling. Here's a quick note about it. This architecture enables multiple CPU sockets on a single motherboard, which means we can plug multiple physical processors on the same motherboard. This is well-suited for workloads demanding high parallelism, such as scientific computing and in-memory databases. Each socket has direct access to its local memory and I/O, but inter-socket communication introduces latency. This is where NUMA-aware scheduling and memory allocation help avoid performance bottlenecks. This setup is adopted when we do not want to manage several physical machines but still need more compute power from a single box. Fascinating stuff.

346

16 Comments

4,000+ Indian Supreme Court cases, visualized on a Qdrant cluster 📊 🚀 40,000+ are queued — this chart will be 10× bigger in a few days 💸 Total infra spend: $0 Here’s how (with a little sarcasm, of course): 1) You don't need LLMs for everything 👨‍💻 - 40,000+ PDFs, very messy and some inconsistent. - I've seen a lot of people opting for LLM or OCR in this case, but that's just to hide their lack of python skills - If you are like me, you'll bang your head on the wall and write python PDF parser without any hacks involved - LLM is only used for 1 in ~1000 spot-checks, not the heavy lifting. Translation: write better code, stop blaming the PDFs. 2) Documentation is underrated 📚 - Every feature you want is one doc page away. - I built clusters, embedding, JSON pipelines, even wired local LLM APIs — all by reading documentation - If reading docs makes you sweat, maybe PM life is calling. - Don’t DM ChatGPT uncle for every minor inconvenience 3) Local models save wallets 💾 - Ollama + OpenHermes-7B (~4GB) handled embedding & indexing. ~8–10M tokens in ~8 hours → $100–$500 in avoided API costs 💰 - If you’ve got no idea how to run local LLMs, either you're ngmi or you better have a $1000 budget for API calls - Even with paid APIs there are rate limits, privacy concerns, etc. I don't like limitations. That's why I am all in on self host - Self-host Chroma if Qdrant Cloud feels pricey. The point ✨ 💪🏻 Skill > spend. Code first. Use LLMs surgically. Ship things 🚢 Thinking of sharing a demo video here. Guess what project I am building. #Qdrant #VectorDB #Ollama #LegalTech #Python

11

After my last post on Speculative Decoding, I kept reading and digging deeper into how researchers are trying to make LLMs even faster. That’s when I came across something called Cache-Enhanced Speculative Decoding. Here’s the problem it tries to solve : In speculative decoding, whenever the draft model predicts wrong, the target model rejects those tokens and all that computation is wasted. Think of it like typing on your phone with autocorrect you type a whole word, it turns out wrong, and you delete the entire thing. Waste of effort. Cache-Enhanced decoding says: “Wait, instead of throwing it all away, why not keep the useful parts in memory and continue from there?” So even when tokens are rejected, the system reuses intermediate results ( cache) instead of recomputing everything. Why does this matter? -> For longer text or reasoning-heavy tasks, this avoids the “wasted effort” problem and makes decoding consistently faster. It’s a simple idea don’t start over every time. Reuse what you can. Sometimes optimization is not about being smarter, but about being less wasteful

14

2 Comments

The tech world and in extension, LinkedIn posts, are now strictly partitioned into 2 camps - AI boomers and AI doomers. AI boomers are the ones who would go on sharing how capable and performant AI is becoming, right from solving Math Olympiad problems to the ability to "reason" through complex problems and express through chain of thought logical deduction, supposedly akin to human thought. In this same camp would be tech leaders and VC investors who are also hyping AI beyond its current capabilities to an unsustainable point. The AI doomers on the other hand are mostly 2 types of people - one who think AI has the capability to do real harm and should be heavily regulated and the others who see the foundation models as just glorified autocomplete. Just like everything else, the truth is almost always somewhere in the middle. LLMs have indeed come a long way from the initial RNNs and serves as a decent competitive threat to search, yet Yann LeCun rightly states that LLMs don't represent intelligence in any sense of the word. AI still has a long way to go for it to be truly meaningful.

14

5 Comments

India just launched its own LLM. Not fine-tuned on someone else’s model. Not a wrapper around GPT. Built entirely from scratch. I was at Bharat Mandapam today when Sarvam AI unveiled India’s first sovereign Large Language Models at the India AI Impact Summit 2026. Here’s what they launched: Sarvam-30B • 16 trillion tokens of pre-training data • 32K context window • Outperforms several well-known global AI systems at its size Sarvam-105B • 128K context window • Zero external data dependency • Competing head to head with GPT-OSS-120B, Qwen3-Next-80B, and GLM-4.5-Air on reasoning benchmarks like GPQA Diamond, AIME 25, and MMLU Pro Both models are open source. The US had its ChatGPT moment in November 2022. China had its DeepSeek moment in January 2025. India just had its Sarvam moment in February 2026. This is not hype. This is a country deciding to stop being an AI consumer and start being an AI creator. Backed by the IndiaAI Mission, trained on Indian languages, Indian scripts, and real world Indian data. If you’re building AI products for India or the Global South, the infrastructure just changed. The foundation is here. What you build on it is up to you. #IndiaAI #SarvamAI #AIImpactSummit2026 #SovereignAI #LLM #OpenSource

24

💼 Layoffs in Tech: It’s Not (Just) AI. Every second post on my feed today talks about layoffs… and understandably so. So far in 2025, we’ve seen over 130,000 tech jobs lost globally. In India alone, the count has crossed 21,000. A few headlines: • Microsoft – 9,000 roles cut • Intel – 24,000 • TCS – 12,000 And many more across both big names and lesser-known companies. And yet, most of the blame seems to be landing on AI and so-called “vibe coding”. Let’s be clear: AI may be reshaping the industry, but it hasn’t matured enough to replace most developer roles yet. This wave of layoffs isn’t just about AI… it’s about: ✅ Overhiring during COVID ✅ Geopolitical & macroeconomic instability ✅ Bloated CTCs and declining productivity ✅ Skill stagnation in a rapidly evolving tech landscape Yes, AI is a force… but it’s not the reason. The tech bubble is simply correcting itself. And it’s a wake-up call. 🧠 Adaptability and upskilling are the real shields in this climate… not fear of being replaced by a bot.

838

29 Comments

If you’re an SDE switching companies in India, Never resign without a concrete offer letter in hand. Not a verbal confirmation. Not a “we’re excited to have you.” A formal letter with role, salary, start date, and location clearly mentioned. I've been a developer since 2011. Switched 7 companies in 14 years. And here's the pattern I follow every time I’ve made a move: If the HR says “Congratulations!” but doesn’t attach an offer letter → Politely thank them, but wait. No documents, no resignation. If they ask for salary slips, increment letters, or your current offer → Share only after you get a formal offer in writing. If you’re pressured to resign before getting confirmation → That’s a red flag. Step back. No matter how exciting the role is. If they refuse to give you a few days beyond notice to join → Think hard. Do they respect your boundaries? If you feel uncertain or uncomfortable during early conversations → Trust your gut. It only gets worse once you're inside. Switching jobs is stressful. Don’t add risk by making assumptions. ✅ Always get the offer letter. ✅ Always confirm your CTC, start date, and notice alignment. ✅ Always walk away from disrespect; no job is worth your dignity and mental health. 

777

42 Comments

The best minds I know in Google, Microsoft or Amazon don’t just prepare for interviews-they build a system for lifelong learning. I was super inspired by Amit Pandey during my work who relentlessly keeps learning more and more! As someone who worked as a Data & Applied Scientist at Microsoft, with a strong track record on Kaggle, CodeForces, and over 10 internships during my college years-I’ve seen one pattern in almost every successful candidate I’ve met: They don’t wait for projects to teach them. They invest in their learning, systematically and consistently. Here’s how you can start building what I call your "Knowledge Portfolio"-something far more powerful than a resume line: 📚 𝗥𝗲𝗮𝗱 𝗮𝘁 𝗹𝗲𝗮𝘀𝘁 𝗼𝗻𝗲 𝘁𝗲𝗰𝗵𝗻𝗶𝗰𝗮𝗹 𝗯𝗼𝗼𝗸 𝗲𝘃𝗲𝗿𝘆 𝗾𝘂𝗮𝗿𝘁𝗲𝗿. Not blogs. Not YouTube summaries. Full books. Start with ones close to your current work, then branch out. Want to go further? Try a book a month. My edge in interviews often came from topics I had read deeply, not just Googled. 🔍 𝗥𝗲𝗮𝗱 𝗻𝗼𝗻-𝘁𝗲𝗰𝗵𝗻𝗶𝗰𝗮𝗹 𝗯𝗼𝗼𝗸𝘀 𝘁𝗼𝗼. FAANG companies don’t hire coders. They hire problem-solvers. And most real-world problems are people problems disguised as tech problems. Learn psychology, product thinking, decision-making. 🎯 𝗘𝘅𝗽𝗲𝗿𝗶𝗺𝗲𝗻𝘁 𝗼𝘂𝘁𝘀𝗶𝗱𝗲 𝘆𝗼𝘂𝗿 𝘀𝘁𝗮𝗰𝗸. If you work in Python, try C++. If you're into REST, explore gRPC. If you use an IDE, try Vim. If you're on Windows, set up Linux at home. Comfort is the enemy of growth. 🌍 𝗚𝗲𝘁 𝗶𝗻𝘃𝗼𝗹𝘃𝗲𝗱 𝗶𝗻 𝗰𝗼𝗺𝗺𝘂𝗻𝗶𝘁𝗶𝗲𝘀. User groups, forums, discussions, side projects-talk to people beyond your company. The best career moves I made came from conversations, not job boards. 🧪 𝗧𝗶𝗻𝗸𝗲𝗿 𝘄𝗶𝘁𝗵 𝗻𝗲𝘄 𝘁𝗲𝗰𝗵 𝗿𝗲𝗴𝘂𝗹𝗮𝗿𝗹𝘆. Even if you never use that framework, language, or library in production, the thinking will carry over. New paradigms fuel better design decisions everywhere. Success in tech isn't about grinding the same skill. It's about compounding your intellectual capital-and letting curiosity lead you to edges others miss. Build your knowledge portfolio like you’re building your future. Because you are.

304

40 Comments

How I reverse two-way door decisions as a Software Engineer at Amazon In a previous post, I talked about bias for action and two-way doors. These are decisions you can reverse if they don’t work out. Shout-out to Abhishek Chaurasiya who reached out via DM and asked me a great follow-up question: “What if you realize right after making a decision that it was the wrong one? What do you do then?” Here’s how I think about it: if you notice the mistake right away, you’re lucky. The longer you go down the wrong path, the more costly it will be to reverse. But the decision to reverse doesn't only affect you, so what do you do? 1. Stop and gather data. Don’t just say “I think this is wrong.” What evidence supports the change? For example, if you decided to make a service synchronous, but then realize the workloads are bursty, gather latency and throughput data to demonstrate why an asynchronous design is more resilient. Evidence will make reversal easier to defend. 2. Anticipate pushback. In a team setting, people may want to stick with the decision. Try to understand their motivations for wanting to keep the wrong decision - some are valid, some aren't (e.g. helps with someone's promotion because a system they own gets more usage). Be ready to explain why pivoting now prevents bigger pain later. 3. Reflect on the root cause. What led to the wrong call? Was it missing data? Rushing under pressure? Missing input from key domain experts? Reflection is how you avoid repeating the same mistake. If it’s your own decision, document the lessons. If stakeholders are involved, bring data and reasoning to the table. Two-way doors only help if you’re willing to step back and learn why you took the wrong turn in the first place. Trying to save face will make the problems worse. How do you handle it when you realize a decision you just made wasn’t the right one? --- Hello, I’m Charles, and I am committed to share my two cents on career-related topics for 100 consecutive days. In these uncertain times, I hope to support those facing layoffs or career challenges. Follow me, and let's navigate this together! (49/100) #BiasForAction #Leadership #DecisionMaking #CareerGrowth #SoftwareEngineering

25

The tech companies who claimed more than 40% code is is being written by AI have resumed the hiring for real engineers. Also now companies have started to fire people in AI divisions as well. The mist is slowly being cleared. AI will mot replace devs. I always advise people to not to be afraid if AI will take thier jobs. I always tell people that AI will create more technical and creative jobs and it will only affect the mundane jobs. AI opens new possibilities of productivity, learning, creativity and prototyping. So people who work by first principles and gain expertise of their domain will never fall short of opportunities. #ai #frontend #backend

38

2 Comments

If you aim for a 1Cr+ role at Meta, you’ll be grilled on system design and DSA If you aim for a 1Cr+ role at Google, you’ll be grilled on system design and DSA If you aim for a 1Cr+ role at Amazon, they’ll again go deep into your system design System design is that one skill that levels up your: – pay scale – responsibilities – impact on the company ➤ Tickets, travel, consumer apps 1. Design an online movie ticket platform like BookMyShow. 2. Design a hotel search and booking engine similar to Booking.com. 3. Design a ride booking platform for cabs. 4. Design a fresh grocery delivery service. 5. Design a food delivery platform. ➤ Social networks and feeds 6. Design a microblogging platform like Twitter. 7. Design a social news feed similar to Facebook. 8. Design a community discussion site like Reddit. 9. Design a nested comments system for posts. 10. Design a voting system for posts and comments. ➤ Media, streaming and content 11. Design a video streaming platform like YouTube or Netflix. 12. Design a live video streaming service for events. 13. Design a podcast hosting and listening platform. 14. Design an image sharing and hosting system like Instagram. 15. Design a movie reviews aggregation service. ➤ Messaging, notifications and collaboration 16. Design a WhatsApp style chat and group messaging system. 17. Design a push notification service for mobile and web. 18. Design a webhook based notification delivery service. 19. Design a tagging or hashtag service for user content. 20. Design a collaborative online document or spreadsheet editor. ➤ Payments, wallets and fintech 21. Design an online payment processing service. 22. Design a digital wallet for peer payments and merchants. 23. Design an ATM and bank account management backend. 24. Design a scheduled digital transaction system for recurring payments. 25. Design an online auction and bidding platform. ➤ Caching, rate limiting and infra 26. Design a web cache service for frequently accessed content. 27. Design an API rate limiting service to protect backend APIs. 28. Design a job or task scheduling system. 29. Design a migration strategy for a large scale caching layer with close to zero downtime. 30. Design a distributed key value store. #sde1 #sde2 #sde3 #sde4 #hiring #jobHunt

383

4 Comments