Poorva Garg
Hi! I am a Ph.D. candidate in the Department of Computer Science at the University of California, Los Angeles, co-advised by Prof. Guy Van den Broeck and Prof. Todd Millstein.
My research lies at the intersection of probabilistic artificial intelligence and programming languages, spanning probabilistic programming languages, formal methods, and large generative models. My work centers on identifying and exploiting structure in probability distributions to design reasoning methods that are scalable, accurate, and widely applicable. This perspective guides my current work on making AI code generation robust and efficient by exposing underlying probability distributions using probabilistic programming. The same perspective earlier led to my work on developing approximate symbolic methods for scalable inference in discrete-continuous probabilistic programs.
Before UCLA, I received my B.Tech. in Computer Science and Engineering from IIT Delhi. Outside of research, I love traveling, writing poems, and socializing.
news
| Nov 1, 2026 | Received the 2027 NUS Computing Ascent Prize. |
|---|---|
| Sep 25, 2026 | CoTs as Probabilistic Programs accepted at NeurIPS 2026. |
| Aug 21, 2026 | Co-organized TPM 2026: The 9th Workshop on Tractable Probabilistic Modeling at UAI 2026 in Amsterdam. |
| Jun 15, 2026 | Started a research internship at the Allen Institute for AI with Kyle Richardson. |
| Apr 20, 2026 | New preprint: Probabilistic Programs of Thought. |
| Oct 12, 2025 | Tuning Random Generators: Property-Based Testing as Probabilistic Programming appeared at OOPSLA 2025. |
| Oct 1, 2025 | Gave an invited talk, A Bitblasting Approach to Probabilistic Programming, at the National University of Singapore. |
selected publications
- Probabilistic Programs of Thought
Poorva Garg, Renato Lui Geh, Daniel Israel, Todd Millstein, Kyle Richardson, Guy Van den Broeck.
arXiv 2026 - Bit Blasting Probabilistic Programs
Poorva Garg, Steven Holtzen, Guy Van den Broeck, Todd Millstein.
PLDI 2024 · Nominated for Distinguished Paper Award - CoTs as Probabilistic Programs: A Programmatic View of Thinking Step-by-Step in Language Models
Kyle Richardson, Yu Feng, Poorva Garg, Junyan Cheng, Dan Roth, Guy Van den Broeck.
NeurIPS 2026 - Tuning Random Generators: Property-Based Testing as Probabilistic Programming
Ryan Tjoa, Poorva Garg, Harrison Goldstein, Todd Millstein, Benjamin C. Pierce, Guy Van den Broeck.
OOPSLA 2025