Utkarsh Tiwari (he/him)

Hi! I'm a 4th year undergraduate studying Computer Science at BITS Pilani. I'm (largely) interested in deep learning and, specifically interpretability and learnability, and more recently systems optimizations for large scale training. In the past I have been involved in computer vision research at the Kiran Bhat Computer Graphics and Computer Vision lab at BITS Pilani and as a summer intern at the STARS team at INRIA Sophia Antipolis where I worked on video understanding models, with a focus on anomaly detection for ego-centric driving scenarios. After which, I worked on interpretability, some ML theory, and formal languages. I interned at Prof. Manik Varma's lab at Microsoft Research, India where I worked on ML Systems and GPU optimizations.

I also enjoy writing blogposts about math and cs when I find the time. Please check them out if you can! More informally, I mantain a list of papers and projects that I've recently read and liked, along with some in-the-works projects of my own. If you find my profile interesting or useful to your projects, or even for just a chat, feel free to reach out. All spelling errors are left in as proofs of human-generated text :)

Email  /  Resume  /  Blogs  /  Scholar  /  Github

Research

Miscellaneous

	Modeling Tool Use in Transformers via Computation Oracles, presented at GenAI Research Group, BITS Pilani Slides
	Talk on Formal Theory for Interpretability at GenAI Research Group, BITS Pilani – Slides, Recording
	We presented our paper as an Oral at the ROAM workshop at ECCV 2024! PPT / PDF
	Some good old natural deduction proofs prepared for a logic course im TAing for PDF
	Photos from ECCV Slideshow Source code adapted from Sankalp Sangle