Software Engineer
LTM [LTI Mindtree] - Jun 2025 → Present
- Automated engineering document processing by architecting a Python-based PDF intelligence pipeline using EasyOCR, Azure AI Document Intelligence, pdfplumber, regex extraction, and LangChain, reducing manual review effort by 90% and enabling processing of 3,000+ engineering PDFs in under 5 hours.
- Built an autonomous procurement agent using LangGraph, CrewAI, FastAPI, React, PostgreSQL, Redis, Celery, MCP, GPT-4o, and Claude, reducing procurement cycle time from 3–8 weeks to < 1 business day while eliminating 50–80+ hours of manual effort per procurement request.
- Engineered resilient multi-agent AI workflows with LangGraph StateGraph, PostgreSQL checkpointing, distributed Celery workers, semantic search via pgvector, and real-time WebSocket streaming to support crash recovery, resumable execution, and scalable enterprise deployment.