Abhinand User

I've built six end-to-end AI applications spanning multi-agent orchestration, hybrid RAG, LLM evaluation infrastructure, and multimodal AI.

In my Agent Eval Platform, I built a multi-tenant system that benchmarks AI agents for correctness, hallucination rate, tool-use compliance, latency, and token cost. I implemented LLM-as-judge evaluation, tool-call auditing, OpenTelemetry tracing, and an async FastAPI, SQLAlchemy, and PostgreSQL backend deployed on Vercel.

I also built a hybrid-RAG CVE intelligence platform that combines dense retrieval, BM25, Reciprocal Rank Fusion, and NVD vulnerability data to generate grounded security analysis reports.

My projects include an autonomous four-agent research pipeline, a multimodal image analysis app, and a local-first PDF RAG CLI with citation-grounded answers. I work across agent design, retrieval pipelines, backend APIs, and cloud deployment using Python, LangChain, LangGraph, FastAPI, and vector databases.