I am a deep learning engineer at NVIDIA, working on frontier model co-design for efficient and agentic inference. Previously, I was an applied researcher at Microsoft Research + M365 Copilot focusing on retrieval, knowledge, and LLM-as-a-judge. I graduated from the University of Michigan in May 2022 with a PhD in machine learning, focusing on graphs, NLP, and Transformers.

See Google Scholar for an up-to-date publications list.

Contact: tarasafavi [at] microsoft.com

Recent news

Misc (updated periodically)