Sotabase

Career

· Research Intern, Google2024–
· Research Intern, Apple2023–
· PhD Student, Carnegie Mellon University2022–
· Master of Language Technologies, Carnegie Mellon University2020–2021
· SDE/Applied Scientist, Amazon2016–2019
· Graduate Student, Carnegie Mellon University
· Bachelor of Technology (B.Tech) in Computer Science (Hons.), National Institute of Technology, Tiruchirappalli
· Visiting Researcher, UC Berkeley

Publications (47)

Politeness Transfer: A Tag and Generate Approach
Annual Meeting of the Association for Computational Linguistics · 2020
169
cited
Rewarding Progress: Scaling Automated Process Verifiers for LLM Reasoning
International Conference on Learning Representations · 2024
167
cited
RL on Incorrect Synthetic Data Scales the Efficiency of LLM Math Reasoning by Eight-Fold
Neural Information Processing Systems · 2024
99
cited
Optimizing Test-Time Compute via Meta Reinforcement Fine-Tuning
International Conference on Machine Learning · 2025
92
cited
Explaining The Efficacy of Counterfactually-Augmented Data
International Conference on Learning Representations · 2020
87
cited
Scaling Test-Time Compute Without Verification or RL is Suboptimal
International Conference on Machine Learning · 2025
70
cited
An Efficient Fault Tolerant Workflow Scheduling Approach using Replication Heuristics and Checkpointing in the Cloud
J. Parallel Distributed Comput. · 2018
41
cited
e3: Learning to Explore Enables Extrapolation of Test-Time Compute for LLMs
arXiv.org · 2025
37
cited
Adversarial Unlearning: Reducing Confidence Along Adversarial Directions
Neural Information Processing Systems · 2022
24
cited
Thinking vs. Doing: Agents that Reason by Scaling Test-Time Interaction
arXiv.org · 2025
21
cited
What Do Learning Dynamics Reveal About Generalization in LLM Reasoning?
arXiv.org · 2024
18
cited
Is Support Set Diversity Necessary for Meta-Learning?
arXiv.org · 2020
17
cited
Two Sides of Meta-Learning Evaluation: In vs. Out of Distribution
Neural Information Processing Systems · 2021
17
cited
Exact Unlearning of Finetuning Data via Model Merging at Scale
arXiv.org · 2025
14
cited
Bitrate-Constrained DRO: Beyond Worst Case Robustness To Unknown Group Shifts
International Conference on Learning Representations · 2023
12
cited
Complementary Benefits of Contrastive Learning and Self-Training Under Distribution Shift
Neural Information Processing Systems · 2023
12
cited
Project and Probe: Sample-Efficient Domain Adaptation by Interpolating Orthogonal Features
arXiv.org · 2023
12
cited
Robust Handwriting Recognition with Limited and Noisy Data
International Conference on Frontiers in Handwriting Recognition · 2020
8
cited
On the Benefits of Public Representations for Private Transfer Learning under Distribution Shift
Neural Information Processing Systems · 2023
6
cited
RLAD: Training LLMs to Discover Abstractions for Solving Reasoning Problems
arXiv.org · 2025
6
cited
Sotabase
Amrith Setlur | Researcher Profile | Sotabase | Sotabase