An AI engineer's shonen story: 3 years of production generative AI — RAG over 50K+ documents, LangGraph agents with guardrails, fine-tuned models on a single A100, and full on-prem MLOps ownership. Every chapter shipped to production.
AI ENGINEER · SAMSUNG ELECTRO-MECHANICS · AUG 2023 — PRESENT
LANGGRAPH · REACT · HUMAN-IN-THE-LOOP
QLORA · PEFT · RAGAS · JENKINS CI
KUBERNETES · HELM · VLLM · MLFLOW
ML training & deployment platform — WPF desktop client + FastAPI backend. Time-to-production: 3 days → under 12 hours across 5 teams. Redis priority job queue running 100 concurrent PyTorch jobs with zero task loss.
Internal document intelligence on LangChain RetrievalQA + Elasticsearch hybrid retrieval — 200+ daily users across multiple engineering teams, deployed on on-prem Kubernetes with PostgreSQL audit logs.
Knox-integrated agent: Planner → Executor → Reviewer graph in LangGraph, conditional retry, human approval on destructive actions. 12 engineers onboard, 4 hrs/week saved on sprint admin.
▲ tap a quest card for the full mission report
▲ tap any item to see how I used it in battle
Samsung Electro-Mechanics Bangalore — chosen from 200+ engineers for delivering the internal LLM-based document search platform.
National-level hackathon at M S Ramaiah Institute of Technology — beat 100+ teams; endorsed by Devfolio & MLH (Nov 2022).
Dayananda Sagar College of Engineering, Bengaluru (2019–2023) — GPA 8.12.
Fastest way to reach me — I actually read my inbox.
dscsmaheshchavan@gmail.com
📞 +91 76766 04917
📍 Bengaluru, India