- 💼 Most recently a Member of Technical Staff at Mercor, building the platform around RL environments and delivering 20+ client projects end to end.
- 🎓 M.S. in Computer Science from the University of Southern California (USC).
- 🏆 Smart India Hackathon 2022 Grand Finale Winner.
- 💬 Ask me about Python, TypeScript, React/Next.js, FastAPI, PostgreSQL, LLM evals, and AWS.
- 📫 Reach me at recruitsaatvik@gmail.com.
Adloom | AI Video Ad Generator (UCLA Gemini API Hackathon)
Generated localized video ads for 3 markets from a single creative brief in 24 hours by chaining Gemini 2.5 Flash, Imagen 3, and Veo 3, with Inngest running the async generation jobs.
Tech Stack: Next.js, TypeScript, Gemini 2.5 Flash, Imagen 3, Veo 3, PostgreSQL, Inngest, Docker.
GitHub · Demo
VOCTV Console | Simulated OCT + Vibrometry Acquisition Console
Desktop console that simulates a Volumetric OCT and Vibrometry system, streaming live B-scans, A-scans, vibration traces, tuning curves, and a 3D basilar-membrane surface at 20 fps. The device runs on a worker thread behind a swappable hardware interface, verified by 34 Qt Test cases in CI.
Tech Stack: C++17, Qt 6, CMake, Qt Test, GitHub Actions.
GitHub
AI FlashCards | Generative AI SaaS
Shipped a generative AI flashcard app that reached 30+ active users and 500+ generated flashcards, with Clerk auth, Stripe payments, and Firebase real-time updates.
Tech Stack: Next.js, LangChain, OpenAI, Tailwind CSS, Clerk, Firebase, Stripe.
Live · GitHub
FitCoach | Microservice Fitness Coaching Platform
Engineered a microservice fitness-coaching platform that broadcasts workout events over RabbitMQ and auto-discovers services with Eureka. It renders live progress in a Next.js dashboard and features a Gemini AI inference layer for personalized coaching. The entire system is orchestrated and deployed end-to-end via Docker on AWS.
Tech Stack: Spring Boot, Next.js, AWS, RabbitMQ, Docker CI/CD, PostgreSQL, MongoDB, Gemini AI.
Xpense | Event-Driven Expense Tracker
Architected an event-driven expense tracker using Spring Boot microservices on Kafka, exposed through a Kong API Gateway and containerized with Docker on AWS. It features a React Native mobile app and a Flask + LangChain service with Mistral AI to convert unstructured notifications into structured JSON, automating 80% of data parsing workflows.
Tech Stack: React Native, Java, Spring Boot, Flask, LangChain, Mistral AI, MySQL, Kafka, Docker, AWS.
