PeopleofSuperintelligence

Researcher

Bhaskar Ramasubramanian

Bhaskar Ramasubramanian is an electrical and computer engineering researcher whose work covers reinforcement learning, language model security and reasoning. After engineering degrees in India, he completed a doctorate at the University of Maryland in 2018 and worked in the University of Washington’s Network Security Lab from 2018 to 2021. He joined Western Washington University in 2021 and coauthored CleanGen in 2024, a method for reducing backdoor attacks on language model outputs. His 2025 work introduced Mix Distillation for small language models, and his 2026 work used multiple training checkpoints to recover reasoning solutions. He is an associate professor at Western Washington, directs its Safe Artificial Intelligence and Cyber-Physical Systems Lab, and holds an affiliate faculty appointment at the University of Washington.

Works
74
Citations
277
h-index
9

Publication numbers from OpenAlex, as of 27 October 2026.

Timeline

  1. 2009–2011

    Engineering degrees in India

    He earns a B.Tech. in electrical and electronics engineering at Visvesvaraya National Institute of Technology in Nagpur in 2009. In 2011, he completes an M.Tech. in systems and control engineering at the Indian Institute of Technology Bombay.

  2. 2016–2018

    A doctorate at Maryland

    At the University of Maryland, College Park, he earns an M.S. in electrical and computer engineering in 2016 and a Ph.D. in 2018. His doctoral work also involves the Institute for Systems Research; Steven I. Marcus and Rance Cleaveland direct his dissertation, Opacity and Structural Resilience in Cyberphysical Systems.

  3. 2018–2021

    Postdoctoral research in Seattle

    In 2018, he becomes a postdoctoral research associate in the University of Washington’s Network Security Lab, a position he holds until 2021. Radha Poovendran and Linda Bushnell host his research in the electrical engineering department.

  4. March 2021

    Reinforcement Learning Beyond Expectation

    In March 2021, he coauthors Reinforcement Learning Beyond Expectation with Luyao Niu, Andrew Clark and Radha Poovendran. The paper develops reinforcement learning algorithms that use cumulative prospect theory to model human preferences and tests them on tasks involving a target and obstacles.

  5. 2021–2025

    Assistant professor at Western Washington

    He joins Western Washington University as an assistant professor in electrical and computer engineering in 2021 and holds this rank until 2025. His research applies control, reinforcement learning, formal methods and game theory to cyber-physical systems.

  6. November 2024

    CleanGen against model backdoors

    At EMNLP in November 2024, he coauthors CleanGen: Mitigating Backdoor Attacks for Generation Tasks in Large Language Models. The method identifies suspicious output tokens and replaces them with tokens from another language model to reduce the generation of content selected by an attacker.

  7. July 2025

    Mix Distillation for small models

    In July 2025, he coauthors Small Models Struggle to Learn from Strong Reasoners in Findings of ACL. The paper introduces Mix Distillation, which combines long and short chains of reasoning to improve the reasoning performance of small language models.

  8. July 2026

    Temporal Sampling for reasoning

    In July 2026, he coauthors Temporal Sampling for Forgotten Reasoning in LLMs at ACL. The study finds that models can lose previously correct solutions during training and introduces a decoding method that draws outputs from multiple training checkpoints to recover such solutions.

  9. As of October 2026

    Associate professor and lab director

    In September 2025, he gains tenure and becomes an associate professor in electrical and computer engineering at Western Washington University. As of October 2026, he directs its Safe Artificial Intelligence and Cyber-Physical Systems Lab and holds an affiliate faculty appointment in electrical and computer engineering at the University of Washington, Seattle.

Career

  1. Since 2025Western Washington UniversityAssociate Professor in electrical and computer engineering
  2. NowWestern Washington UniversityDirector, Safe Artificial Intelligence and Cyber-Physical Systems Lab
  3. NowUniversity of WashingtonAffiliate Faculty in electrical and computer engineering
  4. 2021–2025Western Washington UniversityAssistant Professor in electrical and computer engineering
  5. 2018–2021University of WashingtonPostdoctoral Research Associate, Network Security Lab

Education

  1. 2018University of MarylandPh.D. in electrical and computer engineering
  2. 2016University of MarylandM.S. in electrical and computer engineering
  3. 2011IIT BombayM.Tech. in systems and control engineering
  4. 2009Visvesvaraya National Institute of TechnologyB.Tech. in electrical and electronics engineering

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

Colleagues

People with dated positions at the same organization in the same years.

Sources and credits

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

  1. Bhaskar Ramasubramanian | Electrical and Computer Engineering | Western Washington University eece.wwu.edu. Checked 27 October 2026.
  2. Bhaskar Ramasubramanian | ADAPT adapt.ece.uw.edu. Checked 27 October 2026.
  3. Reinforcement Learning Beyond Expectation arxiv.org. Checked 27 October 2026.
  4. CleanGen: Mitigating Backdoor Attacks for Generation Tasks in Large Language Models - ACL Anthology aclanthology.org. Checked 27 October 2026.
  5. Small Models Struggle to Learn from Strong Reasoners - ACL Anthology aclanthology.org. Checked 27 October 2026.
  6. Temporal Sampling for Forgotten Reasoning in LLMs - ACL Anthology aclanthology.org. Checked 27 October 2026.
  7. Bhaskar Ramasubramanian sites.google.com. Checked 27 October 2026.