Smit Pandit

I’m an aspiring AI/ML engineer focused on building practical, scalable LLM applications—especially RAG pipelines and context-aware systems. I enjoy turning model ideas into reliable software that teams can actually use.

In my Flutter Developer Intern role, I built a scalable POS system using Flutter and Dart, improving billing and order management efficiency. I also integrated GraphQL APIs and used Riverpod to keep the app architecture maintainable and performant.

My project work reflects my AI direction. RepoMind is an end-to-end RAG pipeline for GitHub repositories, featuring language-aware chunking, dual-mode retrieval, and an Ollama-based multi-turn chat flow with prompt engineering. On the backend, I developed a FastAPI service with asynchronous ingestion and efficient query handling.

I also build applied ML products through projects like BreedLens, where I fine-tuned EfficientNet-B3 for dog breed classification and deployed it via a Flask REST API with optimized CPU inference. I’m energized by workflows that combine NLP, retrieval, and deployment—so I’m eager to keep shipping real AI systems end-to-end.