About

I am a fifth-year PhD student in Operations Research at the Department of Management Science and Engineering at Stanford University. I am broadly interested in improving the personalization, safety and alignment of foundation models. I am part of the Stanford Causal AI Lab, advised by Vasilis Syrgkanis. I also collaborate with Andrew Ilyas. My research is generously supported by the William R. and Sara Hart Kimball Stanford Graduate Fellowship.

I interned at Netflix Research in Summer 2026 and 2025, where I worked with Adith Swaminathan, Nathan Kallus, Myungha Jung and Qiuling Xu on topics in personalizing foundation models. Previously I received my master’s degree in Computational Social Science (Economics track) from the University of Chicago and my bachelor’s degree in Mechanical Engineering from the Indian Institute of Technology Madras.

News

  • “Corrupt Plans, Clean Traces” accepted to NeurIPS 2026. See you there!
  • “Pigeonholing” accepted to EMNLP Findings 2026.
  • Started a research internship with the MLIR team at Netflix working on LLM User Simulation.
  • New preprint from my Netflix internship on post-training generative recommenders is out.
  • DPO under unobserved preference heterogeneity accepted to AISTATS 2026.
  • Best Paper Runner-up & Oral at the LM4UC workshop, NAACL.

Research

* denotes equal contribution

Preprints

2026 safety

Implementation Matters in Measuring Chain-of-Thought Monitorability

Matan Shtepel, Keertana Chidambaram, Vasilis Syrgkanis, Andrew Ilyas

Preprint

2026 personalization

Exponential Reward Weighting for Fine-Tuning Generative Recommenders under Sparse and Noisy Feedback

Keertana Chidambaram, Sanath Kumar Krishnamurthy, Qiuling Xu, Ko-Jen Hsiao, Moumita Bhattacharya

Preprint ·arXiv/blog

Publications

2026 safety

Corrupt Plans, Clean Traces: Evading Chain-of-Thought Monitoring with Plan Injection

Keertana Chidambaram, Andrew Ilyas, Vasilis Syrgkanis

NeurIPS 2026 ·arXiv

2026 safety

Pigeonholing: how bad prompts hurt models, causing collapse and mistakes

HyunJi Nam, Keertana Chidambaram, Dorottya Demszky, Natasha Jaques

EMNLP Findings 2026 ·arXiv

2026 personalizationalignment

Direct Preference Optimization With Unobserved Preference Heterogeneity

Keertana Chidambaram, Karthik Vinay Seetharaman, Vasilis Syrgkanis

AISTATS 2026 ·🏆 Best Paper Runner-up & Oral, LM4UC Workshop @ NAACL 2025 ·arXiv

2024 personalization

Personalized Adaptation via In-Context Preference Learning

Allison Lau, Younwoo (Ethan) Choi*, Vahid Balazadeh*, Keertana Chidambaram*, Rahul G Krishnan, Vasilis Syrgkanis

Workshop on Adaptive Foundation Models, NeurIPS 2024 ·arXiv

2024 personalization

Sequential Decision Making with Expert Demonstrations under Unobserved Heterogeneity

Vahid Balazadeh, Keertana Chidambaram, Viet Nguyen, Rahul G Krishnan, Vasilis Syrgkanis

NeurIPS 2024 ·Workshop on AutoRL, ICML, 2025 ·arXiv