AI/ML Engineer specialising in production agentic AI systems, LLM orchestration, and applied ML research.

• 🏛️ Dual degree — IIT Madras (B.Sc Programming & Data Science) + KTU (B.Tech CSE)
• 📄 Published researcher — Springer KAIS Journal + IEEE ICIC3S on fusion GNN-Transformer architectures
• ⚙️ Building production LangGraph multi-agent pipelines at Rappit
• 🔬 3 years experience across agentic AI, RAG systems, and deep learning research
• 📍 Coimbatore, India

stack = {
    "agentic_ai":   ["LangGraph", "Multi-agent orchestration", "HITL gates", "MCP servers"],
    "llm_systems":  ["RAG", "GraphRAG", "LLM evaluation (RAGAS)", "LangSmith"],
    "research":     ["GNNs", "Transformers", "Fusion architectures", "PyTorch"],
    "infra":        ["GCP", "FastAPI", "Docker", "Neo4j", "PostgreSQL", "Redis"],
}

LLM-powered resume optimisation + agentic mock interview system. Live product.
Stack: Python · Flask · Supabase · Vue.js · LLM APIs
🔗 Live demo

Structured financial data extraction from invoice PDFs via LLMs, integrated with Tally ERP.
Stack: FastAPI · Gemini API · SQLite · Vue.js

DDPM (UNet) denoising + CNN regression for target parameter estimation from noisy radar signals.
Stack: PyTorch · Diffusers · scikit-learn

I write about production AI engineering — real benchmarks, real architectures, real failure modes.

• PageIndex vs Naive RAG — benchmark across 58 tests

📖 Read all posts →