Hi, I'm Rohit Punjabi.
Data · Engineering · Product · Claude power user · Open to workI'm an engineer who works right where data, code and users meet. At Shrapnel, a live game studio in Seattle, I built the player telemetry pipeline and the dashboard teams used every week, cutting bug detection and analysis time by ~70%.
Off the clock I'm usually planning my next trip, playing Rocket League, or out with my camera. More about me →
Mumbai, India · Open to remote and relocation · rohitmpunja@gmail.com
I've done all three of these jobs on a real, shipped product, often in the same week. Here's what I'd bring to each one, plus a resume that goes deeper.
I find the answer in messy data and turn it into something people use.
I write production code that ships and keeps working.
I start with the user and measure what changed.
The common thread: I like owning a problem all the way through. Figure out what people need, build the thing, then check the data to see if it worked. Whichever seat I'm in, I bring the other two with me.
Case studies, not repo dumps. Each one starts with a real problem and ends with what changed.
Shows Data · Engineering · Product
Which weapons, maps, and mechanics were unbalancing matches, and where were players hitting bugs we couldn't reproduce?
Live gameplay event streams from thousands of play sessions: kills, weapon usage, movement paths, and power up pickups, collected in engine and stored in a MongoDB cloud backend.
I worked with design and QA to define the metric schema and KPIs, built the in engine collection layer in C++, then led development of a TypeScript and React dashboard so designers could slice sessions without writing queries.
Designers moved from gut feel to data driven balance decisions, engineers found bug conditions from telemetry instead of repro attempts, and the KPIs fed straight into roadmap priorities.
Shows Data · Engineering · Product
What is actually selling, what is dead stock, and how much of each product should the store order in the coming months?
About 50 monthly Excel workbooks spanning 4+ years: inconsistent column names, quantities stored as text, and a product naming change that split sales histories in two.
I built a Pandas pipeline that merges every workbook into one clean table with automated health checks, matches renamed products, and picks and backtests the best forecast per product. Then I wrapped it in a desktop app the team runs themselves, with Claude as my coding partner, turning forecasts into reorder points and a monthly order plan.
The data health checks caught quantities saved as text like "55 PCS", which had silently hidden about 38% of all sales and distorted which products looked like best sellers.
Shows Engineering · Product · Data
Formula 1 data lives in scattered sources: standings in one place, live session data in another, circuit and driver context in a third.
Built a cross platform app (iOS, Android, web) that unifies five free public data sources (Jolpica, OpenF1, Open Meteo weather, Wikipedia, circuit geometry) behind typed API clients with cached queries: live race countdowns, full weekend results, circuit layouts, career stats, and side by side driver comparisons.
Shows Product · Engineering
Reviewers were stuck in a messy, multi step review process. I ran interviews and usability tests, redesigned it into a 2 step flow and built the tool in TypeScript and React, lifting team productivity by 80%.
Shows Engineering
Artists were processing and uploading image assets by hand. I built a Python pipeline on AWS S3 that automated it end to end, cutting manual work by ~30% and removing a whole class of upload errors.
Shows Data · Engineering
I built and trained a convolutional neural network in TensorFlow, then used accuracy and validation curves to diagnose over and underfitting on each iteration.
These are the case studies I would open first. Everything else I have built, custom game engines, simulation and AI demos, the apps I shipped, and two published papers, lives on the archive page. See all projects →
14 months shipping production features and data systems on a live game until a studio wide layoff, then contract work, design freelancing and my own projects since.
One toolkit, used three ways.
I grew up in Mumbai, did my Master's near Seattle, and got my first real engineering job at a game studio, which honestly felt like the dream. These days I'm back in Mumbai. The thing that ties most of my life together is that I love making experiences for people, whether that's a feature, a trip, or a gift.
Travel
Travel is something I genuinely live by. I'm also the go to trip planner in my family: multi city itineraries, bookings, budgets, all of it. It's basically product management with better views.
Games
The Last of Us is my all time favorite. Shaking the controller to fix the flickering flashlight is still one of my favorite bits of game design ever. Right now I'm on Rocket League, Fortnite and F1.
Making things
I design one of a kind magazines for friends and family on special occasions, like an engagement anniversary. Photos, stories, inside jokes, all laid out like a real issue.
Photography
Photography is how I slow down and actually notice what's around me. I post my favorite shots on Instagram at @roviator.rolls →
Sport
I follow football closely and I'm always up for a casual game, or a round of pickleball with friends.
Right now
I'm partway through a 16 week training plan. Same mindset as work: set a goal, track the numbers, show up consistently.
I'm looking for Data Analyst, Software Engineer, and Product Manager roles where I can own problems end to end. I reply fastest by email, and I'm happy to walk through any project in detail.