I am a researcher at Google DeepMind on the GenAI team based in NYC. My research emphasizes principled approaches to understanding and improving deep learning in real-world settings. My current focus is on improving the quality, robustness, and efficiency of reasoning in Large Language Models.
I did my PhD in Machine Learning at Carnegie Mellon University under the supervision of Andrej Risteski and Pradeep Ravikumar. Before entering grad school I worked on Search Research and Machine Intelligence at Google NYC. I received undergraduate degrees in Computer Science and Statistics & Machine Learning from CMU, where my senior thesis was advised by Manuel Blum and Santosh Vempala.
[08/2026] Gemini 3.7 Flash is released! Hopefully more upcoming releases soon 🙂
[04/2026] Our work on the implicit geometric memory of LLMs was accepted to ICML 2026.
[01/2026] Our work on hidden breakthroughs in language model training was accepted to ICLR 2026.
[09/2024] Joined Google.
[05/2024] Our work on one-shot strategic classification was accepted to ICML 2024.
[04/2024] Defended my thesis!
[01/2024] Two papers accepted to ICLR 2024: one on the surprising influence of outliers on neural network optimization and one on provably identifying latent causal factors in representation learning.
Liu Yang (U. Madison → Google DeepMind)
Shikai Qiu (NYU)
Shahriar Noroozizadeh (CMU)
Sara Kangaslahti (Harvard)
Sorawit (James) Saengkyongam (ETH Zürich → Apple)
Sahra Ghalebikesabi (Oxford→ OpenAI)
Bingbin Liu (CMU → Postdoc @ Kempner Institute)
Jivat Neet Kaur (Microsoft Research Fellow → PhD @ Berkeley)
Soundarya Krishnan (Master's @ CMU → Apple)
Deep Karkhanis (Master's @ CMU → Google DeepMind)
Simran Kaur (Bachelor's @ CMU → PhD @ Princeton)
Mariana Tandon (Bachelor's @ Duke → Meta)
For a full list of publications, see my Google Scholar.