I am a Postdoctoral Scholar at Stanford University, co-advised by Scott Linderman and Christopher Ré. I am also an academic partner with Databricks Mosaic AI, where I previously interned. I recently graduated from a PhD at Columbia’s Center for Theoretical Neuroscience, advised by John Cunningham and working closely with Liam Paninski.
I build resource-efficient AI systems for science – in vision, timeseries, and text domains, fusing approaches from statistical ML and CS systems. Most notably, I worked on deep learning systems for tracking animal movement in videos - the Lightning Pose package (Nature Methods, 2024), scalability of Gaussian processes (ICML, 2021), and learning-forgetting tradeoffs in LLM finetuning for math and code generation (TMLR, 2024 (Featured Certification))
Throughout my PhD, I collaborated closely with Lightning AI, named a Lightning Ambassador, and a featured developer in their first DevCon, June 2022.
Here is my CV.
PhD in Computational Neuroscience, 2018-2024
Columbia University
MA in Cognitive Science, 2018
Tel Aviv University
The Adi Lautman Interdisciplinary Program for Outstanding Students (Cog. Sci., Math, Neurobio.), 2013-2017
Tel Aviv University
Compares LoRA versus full-parameter finetuning on challenging code and math tasks; sheds light on the learning-forgetting tradeoffs. Showing that LoRA usually underperforms full finetuning in a new target domain while forgetting less of the source domain.
Introduces a semi-supervised approach to pose estimation, using physically-informed inductive biases to improve generalization with fewer labels. Poses are further refined by combining deep ensembles with state-space models. Open-sourcing a deep learning system that is optimized for efficiency, building on PyTorch Lightning and NVIDIA DALI.
This model disentangles movement that can be quantified by keypoints (e.g., limb position) from subtler feature variations like orofacial movements. We introduce a novel VAE whose latent space decomposes into two orthogonal subspaces – one unsupervised subspace and one supervised subspace linearly predictive of labels (keypoints). The latent space additionally includes a context variable that predicts the video/subject identity.