Research
- Diffusion models for images and language
- Reinforcement learning for LLMs
- Agentic AI
About
I am a third-year PhD student at Texas A&M University, co-advised by Dileep Kalathil and Krishna Narayanan. Before that, I completed my M.Sc. in Algorithms and Computation and my B.Sc. in Computer Engineering at Sharif University of Technology.
News
- Attending COLM 2026.
- Released AgentDiscover, a coding agent that plans its own search for discovery.
- Two papers accepted at NeurIPS 2026.
- Our poster on inference-time search with side information was selected among the best posters in the department.
- Started my PhD at Texas A&M University.
Publications
-
Under review
AgentDiscover: Autonomous Discovery with Minimal Search Scaffolding
A coding agent plans its own search for discovery instead of following a hand-designed algorithm, and matches or beats existing frameworks at lower cost.
-
NeurIPS 2026
Inference-Time Search using Side Information for Diffusion-based Image Reconstruction
A plug-and-play inference-time search that adds side information to pretrained diffusion inverse-problem solvers without retraining.
-
NeurIPS 2026
Reinforcement Learning for Diffusion LLMs with Entropy-Guided Step Selection and Stepwise Advantages
Casts diffusion-LLM denoising as an MDP and focuses policy updates on high-entropy steps, with stepwise advantages and no value network.