Researcher

Karan Singh

Karan Singh is a machine learning theorist and an Assistant Professor of Operations Research at Carnegie Mellon's Tepper School of Business. After a BTech at IIT Kanpur, he did his PhD at Princeton with Elad Hazan, coauthoring work on spectral filtering for linear dynamical systems in 2017 and on online control with adversarial disturbances in 2019. Princeton named him a Jacobus Fellow in 2020, its highest graduate student honor, and Hazan credited his PhD work with laying the foundation of a new sub-discipline called non-stochastic control. Following his PhD in November 2021 and a year as a postdoctoral researcher at Microsoft Research in Redmond, he joined Carnegie Mellon in 2022. His 2026 book with Hazan, Introduction to Online Control, was published by Cambridge University Press.

Works
21
Citations
10
h-index
1
Badges

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. 2015

    Computer science at IIT Kanpur

    He earns a BTech in computer science at the Indian Institute of Technology Kanpur in 2015. He receives the President's Gold Medal for the best academic performance in his graduating class.

  2. 2017

    Spectral filtering for dynamical systems

    With Elad Hazan and Cyril Zhang, he publishes Learning Linear Dynamical Systems via Spectral Filtering at NIPS 2017. The algorithm predicts linear dynamical systems online by convolving a time series with the eigenvectors of a Hankel matrix.

  3. 2019

    Regret bounds for online control

    At ICML 2019, he coauthors Online Control with Adversarial Disturbances with Naman Agarwal, Brian Bullins, Elad Hazan and Sham Kakade, an efficient algorithm with nearly tight regret bounds for linear dynamical systems. Logarithmic Regret for Online Control, his NeurIPS 2019 paper with Agarwal and Hazan, shows that regret can scale as polylog(T), and it wins a best paper award at the NeurIPS 2019 OptRL workshop.

  4. February 2020

    A Jacobus Fellowship

    In February 2020, Princeton names him one of four Jacobus Fellows, its highest graduate student honor. Elad Hazan credits his PhD work with laying the foundation of a new sub-discipline of computer science called non-stochastic control.

  5. 2021

    AI control for ventilators

    In 2021, he is a graduate student on Elad Hazan's Princeton team that applies artificial intelligence to ventilator control during the COVID-19 pandemic, together with Daniel Cohen's mechanical engineering lab. The team's paper, Machine Learning for Mechanical Ventilation Control, appears in the ML4H 2021 workshop track.

  6. November 2021

    A Princeton PhD and Microsoft Research

    He completes his PhD in computer science at Princeton in November 2021, advised by Elad Hazan. He then spends a year as a postdoctoral researcher at Microsoft Research in Redmond.

  7. 2022

    Joining Carnegie Mellon's Tepper School

    He joins Carnegie Mellon's Tepper School of Business in 2022 as Assistant Professor of Operations Research. At NeurIPS 2022, he and coauthors Nataly Brukhim and Elad Hazan present A Boosting Approach to Reinforcement Learning, which reduces reinforcement learning to a sequence of weak learning problems.

  8. As of October 2026

    An online control textbook

    Cambridge University Press publishes Introduction to Online Control in 2026, a tutorial book he writes with Elad Hazan on online nonstochastic control. With Maria-Florina Balcan and Tejas Pagare, he releases Learning to Price with Persuasion in August 2026, a learning-theory model of marketplaces that combines information design and mechanism design. He is an Assistant Professor of Operations Research in the Tepper School of Business at Carnegie Mellon University.

Career

  1. Since 2022Carnegie Mellon UniversityAssistant Professor of Operations Research
  2. From 2021Microsoft ResearchPostdoctoral Researcher

Education

  1. 2021Princeton UniversityPhD in Computer Science
  2. 2015Indian Institute of Technology KanpurBTech

Most-cited works

  1. Introduction to Online Control2026 · Cambridge University Press eBooks · 9 citationsWith Elad Hazan
  2. The Theory of Online Control2025 · Statistical Science · 1 citationsWith Elad Hazan
  3. Background in Control and Reinforcement Learning2026 · Cambridge University Press eBooks · 0 citationsWith Elad Hazan
  4. Basics of Online Control2026 · Cambridge University Press eBooks · 0 citationsWith Elad Hazan
  5. Dynamical Systems2026 · Cambridge University Press eBooks · 0 citationsWith Elad Hazan

Sources and credits

Photo: We have no published photo of Karan Singh yet. Send us a free-licensed photo.

  1. Karan Singh | About | Carnegie Mellon University scholars.cmu.edu. Checked 20 October 2026.
  2. Meet Our New Faculty Members - Tepper School of Business - Carnegie Mellon University cmu.edu. Checked 20 October 2026.
  3. [1711.00946] Learning Linear Dynamical Systems via Spectral Filtering arxiv.org. Checked 20 October 2026.
  4. Online Control with Adversarial Disturbances proceedings.mlr.press. Checked 20 October 2026.
  5. Logarithmic Regret for Online Control papers.nips.cc. Checked 20 October 2026.
  6. Karan Singh i-am-karan-singh.github.io. Checked 20 October 2026.
  7. Four graduate students named Jacobus Fellows - The Princetonian dailyprincetonian.com. Checked 20 October 2026.
  8. Engineering and artificial intelligence combine to safeguard patients’ lives | CS cs.princeton.edu. Checked 20 October 2026.
  9. A Boosting Approach to Reinforcement Learning proceedings.neurips.cc. Checked 20 October 2026.
  10. Introduction to Online Control cambridge.org. Checked 20 October 2026.
  11. [2608.16699] Learning to Price with Persuasion arxiv.org. Checked 20 October 2026.