At Aurora eLABS, I deployed a Small Language Model on a GPU-based cloud server for real-time SMS classification and structured data extraction. I exposed it through FastAPI inference services with asynchronous processing.
I designed a RAG pipeline that ingests documents, retrieves relevant context with MongoDB Atlas Vector Search, and generates LLM responses. I also added metadata-based filtering and hybrid routing between RAG pipelines and SQL systems.
I built AI features including semantic tag matching, an AI Planner, and OCR and call transcription pipelines with LLM-based summarization. I also explored planner–executor workflows using Google ADK concepts.
I developed the Platix mobile application, including OTP auto-read and Google Maps navigation, and deployed it to the Play Store and Apple App Store. Across GME, MarketsAndata, and Happytechies, I took primary responsibility for frontend development and worked on performance, SEO, payments, and email services.
