Jyotirmoy Konwar

I am Jyotirmoy Konwar, an AI/ML engineer and Dual Degree student at Rajiv Gandhi Institute of Petroleum Technology, pursuing a B.Tech in Computer Science and an M.Tech in Artificial Intelligence (2022–2027). I focus on LLMs, multimodal AI, agentic systems, model alignment, and efficient inference.

Most recently I was a Technical Intern at Siemens (Bengaluru), where I built an autonomous agentic code-repair package that fixes CI/CD build failures with verifiable patches, and an LLM evaluation framework for legacy code migration agents. Before that I worked as an MLOps Intern at Borde, did research internships at IIT Jodhpur and IIT Guwahati, and shipped a quantized multimodal document-parsing pipeline for a freelance client.

My projects span RLHF/GRPO fine-tuning (teaching Gemma-3 to reason), LLM routing for coding agents, secure RAG, graph learning, and diffusion language models. I won the Siemens Aavishkar Hackathon 2026, was a Grand Finalist at the Meta PyTorch OpenEnv Hackathon, and ranked 52nd in the Amazon ML Challenge 2025.

I enjoy taking ideas end to end, from research and modeling through deployment, evaluation, and MLOps.