I am a Staff Research Scientist at Google Research. I am excited about structured data foundation models and making + deploying them in real enterprise problems. Nowadays I spend most of my time thinking about the impending seismic shift in structured data prediction pipelines. I am one of the creators of TimesFM and also the research lead of TabFM.
I obtained my PhD from UT Austin in 2019 where I was fortunate to be advised by Dr. Sanjay Shakkottai. I also had the pleasure of working closely with Dr. Alex Dimakis, during my time at UT. I have worked on contextual bandits, extreme classification and black-box optimization.
Updates
[Sep 24, 2026] Our paper Rethinking Post-Training Recipes for Multimodal Time-Series Forecasting, on post-training LLMs to revise TimesFM forecasts using multimodal context, has been accepted to NeurIPS 2026.
[Aug 31, 2026] We released TimesFM-3, our next-generation zero-shot foundation model natively pre-trained for multivariate time-series forecasting.
[Jun 30, 2026] We released TabFM, a zero-shot foundation model for tabular data classification and regression.
[Apr 30, 2026] Our paper Rethinking Multimodal Time-Series Forecasting Evaluation, introducing the TimesX multimodal time-series forecasting benchmark, has been accepted to ICML 2026.