About
Experience & Education
Other similar profiles
-
Sirshendu Maiti
Sirshendu Maiti
Data Science & AI Leader with experience in building and managing high-performing data science teams at Google, Facebook, eBay. <br><br>Currently leading a global team of data science managers, staff/senior data scientists to support Measurement (Experimentation) & Evaluations (Human raters, AI raters), Metrics Research and Product development (Quality and UI evals) for three critical product areas of Google Search & AI Mode, interfacing with 800+ Product Managers, Engineers & cross-functional Partners.<br>Product areas:<br>- Agentic Search (0-to-1)<br>- Real World Journeys (Local & Travel Search verticals)<br>- Next generation of AI-driven Google Finance <br><br>Proven leadership experience in Search & AI Assistant Metrics research & development (Ranking, Retrieval, Relevance, Quality, Entropy), Human/Hybrid Evaluation Methods (Helpfulness, Factuality, Relevance, Grounded-ness, Freshness), Voice assistant & conversational AI, Online experimentation (sequential and bayesian A/B testing, multi-arm bandit), Causal inference, Advanced machine learning, Generative AI (prompt-engineering, fine-tuning).<br><br>Specialties: Conversational AI Metrics, Generative AI, Machine Learning, Human/Hybrid Evaluation Methods, NLP, Search/Ranking Metrics, Monte Carlo Simulation, Causal Inference, Stochastic Models, Survival Models, Bayesian Experimentation, Credit Scoring, P&L Forecasting.
Explore more posts
-
Dzuy Linh
I’ve finally figured out a pretty awesome workflow to quickly build things. Use all the AIs, folks. 1. Talk to humans about the problem 2. Talk to Claude about said problem. Strategize. Claude’s reasoning and practicality is a good foundation. Output something PRD-like so you have a North Star. Create a single large prompt to build a prototype. 3. Paste prompt to Lovable and build out the UI. Fine tune in Lovable because it's fast and has the best understanding of good UI patterns. No backend, no auth, no integrations yet. Just clickable prototype to get human feedback. 4. Setup a backend and a frontend somewhere (go, supabase, railway, vercel, etc). 5. Open the Lovable code in Codex to build out the BE/FE and tell it to clean up Lovable’s crappy code. 6. Finish off in Codex with integrations, auth, security checks Iterate / ship like crazy Human —> Claude —> Lovable —> Human —> Codex <—> Ship Funny part is once you adjust to this new speed, you only want to go faster and non-stop. For those who have gone deep into AI builder mode, what’s your workflow look like? I can’t be the only one suddenly looking to shave off minutes while I tell the bots what to do.
4 Comments -
Sameer Bhardwaj
You are in a system design interview at Meta (which owns WhatsApp) for the IC4 role The interviewer leans in and asks: "Why can I log into Discord or Slack on a new laptop and instantly see my old chats, but WhatsApp needs a backup or a long 'restoring history' step? I asked you to design a chat application. How would these two approaches change your system?" Here is how you should break it down 👇 Both WhatsApp and Discord look like simple chat apps. Under the hood, they are built on two very different philosophies. Btw, we’ve recently launched our system design mock interview tool at Layrs. If you’re preparing for a switch or SD interviews, it’s free to use: http://layrs.me/interviews 1. WhatsApp – Phone is the source of truth Idea: Your device is the real database. The server is mostly a courier. What happens when you send a message: – You type the message. – It is encrypted on your phone with keys stored only on devices in that chat. – WhatsApp server receives the encrypted blob, stores it briefly, and forwards it to the receiver. – Once all devices in that chat acknowledge, the server is allowed to drop the message content. – Your long-term history lives in a local database on phone (SQLite style), plus whatever you back up to iCloud or Google Drive. What this means: – New phone, no backup = almost no history. – Multi device is hard, because every device needs keys and a copy of messages. – Server cost is lower, because they are not keeping infinite message history for every user. – Privacy is strong. Even WhatsApp cannot easily scan your chat content, because they do not keep decrypted data. In your design answer, you can mention: + Local storage on device + End to end encryption + Store and forward model + Optional cloud backup handled as a separate flow 2. Discord or Slack – Server is the source of truth Idea: Servers hold a permanent log of messages. Your device is just a window. What happens when you send a message: – You hit send. The message goes to a backend service. – Backend writes it into a big distributed database or log store (for example Cassandra or ScyllaDB). – Indexes and partitions are based on channel, user, time, workspace, so reads scale horizontally. – Any device that logs in simply fetches messages from the server and renders them. What this means: – New laptop, new browser, office desktop, phone, all show the same history instantly. – Multi device is easy, because all devices just read from the same central store. – Server cost is higher. You are paying to store years of chat history and to replicate it. – Privacy tradeoff. Company can run search, analytics, and moderation on server side data. In your design answer, you can mention: + Distributed database for messages + Append only style message log + Search indexes, pagination, and caching + Stronger focus on availability and horizontal scalability Check my 1st comment as well:
29 Comments -
Sanchit Narula
The best advice I got on my promotions as a software engineer was from a Principal Architect at Amazon after he told me his story of getting promoted to Principal, and this has stayed with me my whole career. He said, very simply: > Your architecture can be a love letter to your technical brilliance, > but if it does not move customer and company goals, it will not move your level. That hurt a little. Because I recognised myself in it. A few years ago I wanted to design something with every shiny thing in the book: API Gateway, Lambda, DynamoDB, event bus, the works. On paper it was beautiful. In reality it was expensive, slow to ship, and did not match the skills of the team that would maintain it. So we shipped a hybrid architecture instead. Fewer moving parts, more reuse of existing services, cheaper to run, easier to hand over. Over time, here is how my thinking about promotions changed. 1. Start from the business goal, not the tech stack Before talking about Kafka vs SQS or MySQL vs DynamoDB, answer this first: What money, risk, or customer pain will this remove? Promotions track that, not the number of services in your diagram. 2. Ship finished products, not clever demos Leaders remember the feature that went live, stayed stable, and earned or saved real money. Drive the entire lifecycle yourself: design, build, launch, oncall, cleanup, docs. 3. Make the team faster, not just yourself Tools, scripts, templates, better CI, better dashboards. If people quietly say “things move faster when you are on this” your manager already has half a promotion case. 4. Do your share of the unsexy work Oncall, migrations, bugs, incident reviews. Senior folks who avoid this slow the entire org. Senior folks who lean into it become trusted very quickly. 5. Grow people, not just systems Mentor juniors, run design reviews, write clear docs, set standards. When your manager sees that you are already acting like the next level, formal promotion becomes a catch up, not a favour. 6. Be proactive about your promotion, not obsessed with it Have periodic one to ones. Talk about where you want to go and ask what evidence your manager would need to support that. Remember your manager has ten other people to manage. It is your responsibility to make your impact visible. Do not be desperate. Understand the timelines and work within them. As you get close to the boundary, collect the missing data points: impact metrics, incident write ups, design docs, peer feedback. Make the story easy to tell. Most importantly, work for the business, not for promotion-driven development. If your roadmap only exists to tick boxes on a promotion rubric, people can feel it. If your roadmap clearly creates value, saves cost, or reduces risk, promotion becomes the natural side effect. So instead of asking “What new tech can I use this year?” Ask, “Where can I create obvious business impact and leave the system and the people around me better than I found them?”
15 Comments
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top contentOthers named Chris Nelson
3066 others named Chris Nelson are on LinkedIn
See others named Chris Nelson