Zeeshan Patel

I am an undergraduate student at UC Berkeley studying Computer Science and Statistics. My main research focuses are in generative models and AI understanding. I am advised by Professor Alexei Efros and Yossi Gandelsman at Berkeley Artificial Intelligence Research (BAIR).

Currently, I am interning as a Deep Learning Algorithms Engineer at NVIDIA NeMo. Previously, I've interned as a ML Engineer at Apple AI/ML within the Information Intelligence team, focusing on foundation models, and at Verkada as a ML Engineer on the Special Projects team.

Email  /  Github

I'm broadly interested in machine learning, computer vision, and generative AI. Specifically, I'm interested in the mechanisitic interpretability of deep learning models and how to create AI systems that can generalize under distribution shifts. I'm also curious about the intersection of vision and language and how visual systems can leverage language to interact with humans.

We introduce Storytelling With Action Guidance (SWAG), a novel approach to storytelling with LLMs. Our approach reduces story writing to a search problem through a two-model feedback loop. SWAG can substantially outperform previous end-to-end story generation techniques when evaluated by GPT-4 and through human evaluation, and our SWAG pipeline using only open-source models surpasses GPT-3.5- Turbo.

A self-supervised approach for fine-tuning image superresolution models to adapt to new test distributions on-the-fly.