Namasivaayam L

Idea-Code (Autonomous PDLC): Architected a LangGraph Deepagent pipeline using Azure/GitHub MCP servers to automate the PDLC from concept to GitHub PRs. Orchestrated the autonomous generation of BRDs, architectures, and ADO tasks, streamlining the transition from requirements to production code. In A Box (Multi-Agent System): Automated intricate multi-document verification across 50+ checkpoints by implementing a scalable agent system with Google ADK, achieving a 67% reduction in manual processing time. PyroAI Docs (GenAI Extraction): Elevated data extraction accuracy to 90% (from 9%), enabling straight-through processing for over 15 financial document types by engineering robust production prompt pipelines. AI Guardian (LLM Validation): Spearheaded the design and deployment of an ML validation framework to benchmark production LLM models against market competitors, architecting a scalable backend with FastAPI and containerization (Docker/GKE).

Secured the full-time implementation roadmap by demonstrating a 10x accuracy improvement (9% to 90%) via an image-based Gemini extraction Proof of Concept (POC), superseding traditional Deep Learning (DL) methods.

Relevant Coursework: Machine Learning, Deep Learning, Data Science, IoT, Data Analytics, Data Visz., Cloud Computing. Selected as Best Outstanding Student (2024) for maintaining an 8.49 CGPA while directing events and operations for the Microsoft Campus Club and Consumer Club.