Tanishq Sangwan - I build for
millions.
Open source contributor at the Internet Archive, improving retention for 15M+ readers. B.Tech in AI & Data Science at IIT Patna.
AI & Data Science student at IIT Patna, contributing to systems that serve millions of readers.
Moving every two years growing up taught me to actually live somewhere, not just pass through - five years of French and Japanese, state-level tennis, district-level swimming, family hikes and cave exploring. I try to bring that same instinct to new codebases and teams.
It started with a "Learn to Code in Python" book, cracked open in a boring IT class back in 2019 - and it gave me a strange sense of peace. With everything else changing every two years, code became the one constant. Since then I've fallen in love with full-stack development - there's something mesmerizing about owning every layer of a platform, from backends in C++, Python, and Node.js, to database design in MongoDB and PostgreSQL, to the frontend in React that people actually touch. But what drives me more than the tech itself is who it's for: leading development at ZNotes, an educational platform working to close the global learning gap, for 100,000+ students taught me how much more I care about coding with purpose than coding for its own sake. That same instinct pulled me into open source - first contributing to OpenMRS's medical record systems for hospitals, and now with the Internet Archive's Open Library, where I joined through Google Summer of Code and I'm still contributing today, focused on improving retention for a platform 15 million people actually use.
Academics
Current Focus
Languages
Frameworks & Libs
Databases & Cache
Infra & Tooling
Google Summer of Code - Software Engineer
- Designed production A/B testing framework with probabilistic bucketing and Matomo analytics - eliminated layout flickering for 15M+ patrons.
- Reconciled and deduplicated three disparate data streams (active CDL loans, PostgreSQL records, client telemetry) - unified reading dashboard across the platform.
- Built Solr-powered discovery and recommendation features backed by Memcached caching - served low-latency suggestions across 7M+ catalog books.
- Engineered Python server-side batch handler replacing legacy client-side Goodreads import loop - reduced network overhead and import latency.
- Continued post-program contribution - currently leading reader retention and onboarding initiatives.
Software Engineer → Technology Lead
- Re-architected monolithic services onto AWS serverless (Lambda, EC2, S3) - cut API latency 40%, infra cost 60%, for 100k+ MAU.
- Architected distributed FastAPI and Node.js REST APIs with MongoDB pipelines - enforced strict JWT-based RBAC for secure data access.
- Containerized backend services with Docker and engineered server-side PDF generation via webhooks - reduced publishing turnaround by 75%.
- Led team of 5 engineers through agile sprints and reviewed 100+ PRs - optimized student engagement features based on telemetry.
VaultPay
A complete banking application built with Next.js and MongoDB that handles transfers, loans, KYC, and card issuance. Uses ACID transactions across replica sets so ledger balances stay reliable under concurrent operations.
The Unspoken Project
An infinite virtual canvas where anyone can pin anonymous notes paired with emotion emojis. Built with React and Konva.js so users can pan and zoom across a massive 4000x4000 coordinate space without render lag.
Velora
Built the entire frontend and backend platform for Velora, an AI-powered market intelligence system awarded Silver Medal at Inter IIT Tech Meet 14.0. Focused on low-latency WebSocket streaming, Kafka consumers, and the interactive Next.js trading dashboard.
Hoopla
A clean, Notion-like to-do application built with Next.js and MongoDB. Created as a distraction-free workspace for organizing daily sprints and task notes without bloated UI.