Sotabase

Career

· PhD Student, Carnegie Mellon University2023–
· MS in Computer Science, Stanford University2022–2023
· Freelance Software Engineer, Upwork2021–
· BS in Mathematics (with distinction), Stanford University2019–2022
· Software Development Intern, WYFT Inc.2019–2019
· C Prescott Davis Scholar, Columbia University in the City of New York2017–2021
· Chief Academic Coordinator, Bangladesh Physics Olympiad2014–2017
· PhD Student, CMU

Publications (29)

Surgical Fine-Tuning Improves Adaptation to Distribution Shifts
International Conference on Learning Representations · 2022
259
cited
Preference Fine-Tuning of LLMs Should Leverage Suboptimal, On-Policy Data
International Conference on Machine Learning · 2024
176
cited
Scalable deep learning to identify brick kilns and aid regulatory capacity
Proceedings of the National Academy of Sciences of the United States of America · 2021
48
cited
No True State-of-the-Art? OOD Detection Methods are Inconsistent across Datasets
arXiv.org · 2021
33
cited
Can Large Reasoning Models Self-Train?
arXiv.org · 2025
24
cited
When to Ask for Help: Proactive Interventions in Autonomous Reinforcement Learning
Neural Information Processing Systems · 2022
21
cited
Do Deep Networks Transfer Invariances Across Classes?
International Conference on Learning Representations · 2022
19
cited
Training a Generally Curious Agent
International Conference on Machine Learning · 2025
12
cited
Reasoning as an Adaptive Defense for Safety
arXiv.org · 2025
10
cited
Conservative Prediction via Data-Driven Confidence Minimization
Trans. Mach. Learn. Res. · 2023
7
cited
Offline Retraining for Online RL: Decoupled Policy Learning to Mitigate Exploration Bias
arXiv.org · 2023
4
cited
State Combinatorial Generalization In Decision Making With Conditional Diffusion Models
Trans. Mach. Learn. Res. · 2025
2
cited
Use of bidirectional DC-DC Buck Boost Converter in Distributed Energy Resources (DER)
2017
2
cited
Accelerating Diffusion Planners in Offline RL via Reward-Aware Consistency Trajectory Distillation
2025
1
cited
Self-Regulation and Requesting Interventions
arXiv.org · 2025
1
cited
Accelerating Diffusion Models in Offline RL via Reward-Aware Consistency Trajectory Distillation
arXiv.org · 2025
DO DEEP NETWORKS TRANSFER INVARIANCES
2022
Efficiently Satisfying Subgroup Fairness in Generalized Classification Settings
2019
Expanding the Capabilities of Reinforcement Learning via Text Feedback
2026
Fine-tuning LLM Agents with Retrospective In-Context Online Learning
Sotabase