Scholar
Kavosh Asadi
Google Scholar ID: -2qyBJEAAAAJ
Meta
Reinforcement Learning
AI Alignment
Optimization
Follow
Homepage
↗
Google Scholar
↗
Citations & Impact
All-time
Citations
3,027
H-index
15
i10-index
19
Publications
20
Co-authors
20
list available
Contact
CV
Open ↗
Twitter
Open ↗
GitHub
Open ↗
LinkedIn
Open ↗
Publications
7 items
Generative Reasoning Re-ranker
2026
Cited
0
Displacement-Resistant Extensions of DPO with Nonconvex $f$-Divergences
2026
Cited
0
Structure Enables Effective Self-Localization of Errors in LLMs
2026
Cited
0
Learning to Reason Efficiently with Discounted Reinforcement Learning
2025
Cited
0
InstructVTON: Optimal Auto-Masking and Natural-Language-Guided Interactive Style Control for Inpainting-Based Virtual Try-On
2025
Cited
0
C-3DPO: Constrained Controlled Classification for Direct Preference Optimization
2025
Cited
0
Adjoint sharding for very long context training of state space models
2025
Cited
0
Resume (English only)
Academic Achievements
Published over a dozen papers at top-tier AI conferences including NeurIPS, ICML, ICLR, AAAI, and ACL
ICML 2024: Paper on learning the target network in function space accepted
ICLR 2024: Paper on foundation models for continual learning accepted
NeurIPS 2023: Two papers accepted
RLC (first Reinforcement Learning Conference): Paper on fairness in RL accepted
NeurIPS 2022: Two papers accepted
AAAI 2021: Two papers accepted
NeurIPS 2021: One paper accepted
AISTATS 2022: One paper accepted
Co-authored the RL chapter in the D2L (Dive into Deep Learning) book
Research Experience
2025–Present: Senior Scientist at Meta’s RL team
2020–2025: Scientist at Amazon (Senior from 2024)
Gave a talk at Seattle Mind and Machines meetup (UW)
Gave a talk at Amazon’s RL reading group
Guest lecturer for Harvard’s ML class
Background
AI scientist aiming to understand the computational principles underlying intelligence
Focuses on agents that interact with sequential environments and improve behavior through trial and error—i.e., the reinforcement learning problem
Academically interested in the optimization problem in value function learning
Applies research to developing assistive AI agents that interact with humans and learn from feedback
Aspires to build AI agents that co-exist with humans and help them live their best lives
Miscellany
Moved from Oahu, Hawaii to Seattle, WA
Moved from SF Bay Area to work remotely from Hawaii
Moved from Providence, RI to SF Bay Area
Currently based in the Bay Area working at Meta
Open to connecting with AI scientists, engineers, and students via email (firstname@alumni.brown.edu)
Co-authors
20 total
Michael Littman
Brown University
Alex Smola
Boson AI
George Konidaris
Brown
Rasool Fakoor
Amazon Web Services
Dipendra Misra
Staff Research Scientist, Mosaic Team, Databricks
Co-author 6
David Abel
DeepMind / University of Edinburgh
Shoham Sabach
Associate Professor, Cornell, School of Operations Research and Information Engineering
×
Welcome back
Sign in to Agora
Welcome back! Please sign in to continue.
Email address
Password
Forgot password?
Continue
Do not have an account?
Sign up