Senior LLM Agent Engineer

Shivam Mittal is an Applied Scientist with 4 years of experience at Microsoft, specializing in end-to-end LLM agent development, including synthetic data generation, SFT/RLHF post-training, and evaluation infrastructure. He has significantly improved agent performance (e.g., Computer Use Agent task pass rate 45%
78%) and developed an LLM evaluation platform serving 500+ Microsoft teams. Proficient in Python, PyTorch, advanced LLM post-training, RAG, and agent evaluation. He is seeking Founding AI Engineer/Applied AI Engineer roles at Seed–Series A AI startups, focusing on coding agents, reliability, and evaluation infra. He is a Kaggle x3 Expert and a top performer in Amazon ML Challenges. Based in India, he is open to remote work with US overlap and willing to relocate to the US.