Rhythm Syed
rhythm dot syed at columbia dot edu

I am a PhD student in Computer Science at Columbia University, happily advised by Yunzhu Li and Tony Dear. I received my B.S. in Computer Engineering from Purdue University and M.S. in Computer Science from Georgia Tech.

I am interested in world models and vision-language-action policies for robotic manipulation, specifically building generalist robots that can reason about and interact with the physical world.

Google Scholar  /  LinkedIn  /  X  /  GitHub

profile photo

News

[Apr 2026] Our paper, Interactive World Simulator, was accepted to RSS 2026!
[Apr 2026] Presented at Amazon Research Day in Palo Alto.
[Mar 2026] Gave a talk at the Computing Research Association Grad Cohort Workshop in Seattle.
[Jan 2026] Started my internship at Toyota Research Institute on the Large Behavior Models team.
[Nov 2025] Serving as a reviewer for IEEE Robotics and Automation Letters.
[Jan 2025] Started my PhD at Columbia University!
[Sep 2023] Started teaching NLP for Columbia's AI Executive Education program.
[Oct 2023] Excited to be an instructor for Correlation One's Data Science For All Fellowship!
[Aug 2021] Joined the Verizon AI Center as a Senior AI Scientist.
[Aug 2021] Graduated with an M.S. in Computer Science from Georgia Tech.
[May 2021] Presented our work on ontology learning and knowledge graph construction at the GTRI Research Conference.
[Jan 2021] Released our paper on latency prediction for Neural Architecture Search.
[Aug 2020] Started research at Georgia Tech Research Institute with Santiago Balestrini and Chao Zhang.
[Jul 2019] Joined AT&T Labs as a Data Scientist.
[May 2019] Graduated with a B.S. in Computer Engineering from Purdue University.
[Nov 2018] Presented our Indoor Positioning System at the Purdue Undergraduate Research Exposition.

Research
Interactive World Simulator for Robot Policy Training and Evaluation
Yixuan Wang, Rhythm Syed, Fangyu Wu, Mengchao Zhang, Aykut Onol, Jose Barreiros, Hooshang Nayyeri, Tony Dear, Huan Zhang, Yunzhu Li
RSS 2026
project website / arXiv / code / models & data

We introduce a world model that runs at 15 FPS for over 10 minutes on a single RTX 4090 GPU and predicts photorealistic and physically accurate future frames. Our model is interactive in real time and performs action-conditioned video prediction which enables high quality data generation for policy training and evaluation making it scalable, reproducible, and faithful.

Teaching & Service

I enjoy teaching and mentoring - please feel free to reach out!

COMS 4705 NLP Adjunct Lecturer: Columbia AI Executive Education
Instructor: Correlation One's Data Science For All Fellowship
Teaching Assistant: NLP @ Georgia Tech, Electronic Devices and Design @ Purdue
Cybersecurity Mentor: GenCyber Summer Camp @ The University of Memphis
Journal Reviewer: IEEE Robotics and Automation Letters, IEEE Transactions on Robot Learning

Visitor map

Last updated May 2026. Template from here.