Hey, I’m Het,
Associate Software Engineer

I build production RAG and agentic AI systems. My stack: Python, FastAPI, LangChain, LangGraph, Pinecone, FAISS, LlamaIndex — deployed on Azure with Docker and CI/CD.
 

SourcePro — Built a RAG system over full ERP documentation, including image retrieval without multimodal models to keep costs down. Also built a SQL agent over a database with 800+ tables, handling natural language queries at under 500–800ms response time.
 

GrowthHive — Built a lead qualification agent that engages prospects, scores them against criteria, and auto-schedules meetings end-to-end.
 

Noodle AI — Built multiple LangChain-based services, each running a single-purpose agent for evaluating user intent and responses during AI-driven interviews.
 

TeachBetter AI — Worked on Concept Mastery, a tool that generates structured learning content for any subject at a given standard and issues certifications after assessment.

Strong on retrieval architecture, agent workflow design, failure handling, and token cost control. All systems shipped to production.