At isaii.in, I developed object-oriented RESTful APIs with Express.js and MongoDB for product catalog management, improving product loading speed by 15%. I also integrated APIs to synchronize product listings and orders across web and mobile platforms.
For the Redrob Intelligent Candidate Discovery Challenge, I built an AI-powered candidate-ranking system that processed 100,000 profiles using hybrid retrieval, semantic search, and behavioral signals. I developed a ranking pipeline combining BM25, embedding similarity, feature engineering, and LLM-assisted reasoning under CPU-only, offline, and runtime constraints.
My projects include Whistle, a communication and workspace application with real-time messaging, and a rail complaint system for semantic analysis and tracking. I’ve also contributed to SurfSense and Sugar Labs, adding configurable LLM settings in one project and removing deprecated caching code in the other.

