At Novel Office India, I engineer Python and FastAPI solutions for enterprise voice, document, and data-processing workflows. I apply GPT-4o, LLaMA3, and Qwen3 to tasks including RAG, summarization, ranking, and question answering.
For SourcingAI, I built a pipeline to ingest, validate, score, and prioritize more than 300 real-estate properties and their associated content. Its XGBoost scoring engine classifies leads into seven priority tiers, while Qwen3 generates structured investment analysis.
I also developed a real-time voice assistant combining Whisper, GPT-4o, text-to-speech, FastAPI, and WebSockets. Its RAG pipelines use FAISS and Chroma to answer questions from internal knowledge, and latency was reduced by approximately 30%.
My other projects include an AI resume-ranking workflow and an LLM-based summarization system for project daily logs. I hold a B.Tech in Computer Science Engineering from MahatmaGandhi University and have certifications in data science with Python and SQL.
