Harini Jeyashree

I build AI systems that actually work in production — not just demos. As a final-year B.Tech IT student specializing in AI/ML Engineering and Generative AI, I've shipped multi-agent LangGraph pipelines, dual-LLM orchestration layers, real-time full-stack applications, and an ML model that hit 95.3% F1 after diagnosing a silent accuracy paradox most engineers would have shipped as a win. My flagship project, Minerva Analyst, is a secure enterprise BI portal with a 4-agent LangGraph pipeline, sandboxed code execution, cryptographic insight snapshots, and fault-tolerant LLM fallback routing — built because most BI tools send your data to the cloud. Mine doesn't. What I bring to a team: → Production-grade GenAI development (LangGraph, LangChain, RAG pipelines, prompt engineering) → End-to-end ML pipelines from messy data to deployed API (Python, FastAPI, Scikit-learn) → Full-stack AI applications (React, Node.js, PostgreSQL, Supabase) → Data analytics and visualization (Power BI, Tableau, Microsoft Fabric) I don't just run models — I design for edge cases, engineer for failure, and document every decision. Currently seeking AI/ML Engineering and Generative AI internships, with full-time availability from 2027. Open to on-site (Chennai) and remote opportunities across world. Let's build something worth shipping. → harinijeyashree878@gmail.com