ML / AI Platform Engineer — continuous LoRA adaptation, LLMOps gateways, LangGraph agents, GraphRAG pipelines, production FastAPI systems.
B.Tech CSE (AI & ML) @ Sushant University · Gurgaon, India
📄 Resume (PDF) · LinkedIn · hansitvarshney@gmail.com
| Project | Description | Links |
|---|---|---|
| AdaptLoop | Continuous LoRA adaptation — Redpanda/Bytewax signals → curated datasets → PEFT → LLM-judge + lag gates → sticky canary into LLMOps gateway | Repo |
| LLMOps Gateway | Async multi-tenant LLM proxy — dual-layer semantic caching, adapter routing + canary, circuit breakers, OpenTelemetry, 127+ tests | Repo |
| EPCDash | Enterprise EPC command center — LangGraph multi-agent ingestion, cash-flow auditing, deployed at SKIC Pvt. Ltd. | Repo · Live demo |
| AEVAR | GraphRAG financial audit pipeline — Pydantic-gated ingestion, Neo4j context, Gemini executive briefings | Repo · Live demo |
Portfolio loop: EPCDash (domain) → AEVAR (knowledge) → LLMOps (serve) → AdaptLoop (adapt)
Python · FastAPI · PEFT / LoRA · MLflow · Redpanda · Bytewax · LangGraph · RAG / GraphRAG · Redis · PostgreSQL · Neo4j · Qdrant · MinIO · Docker · Next.js · Streamlit