Cenk Baykal

Research Scientist, Google Research

Efficient ML systems and algorithms.

I'm a Research Scientist at Google Research. My work focuses on efficient ML, especially algorithms for large-scale learning systems, transformers, distillation, sparsity, and model compression. Before returning to Google, I was a quantitative researcher at Two Sigma. I received my Ph.D. from MIT CSAIL, advised by Professor Daniela Rus.

Portrait of Cenk Baykal

Research

Selected Publications & Preprints

Selected work on efficient transformers, data-efficient distillation, sparsity, model compression, and robotics.

Full publication list on Google Scholar.

Efficient Transformers & Distillation

  1. Cenk Baykal, Dylan Cutler, Nishanth Dikkala, Nikhil Ghosh, Rina Panigrahy, Xin Wang

    Alternating Updates for Efficient Transformers

    NeurIPS 2023 Spotlight
    Read on arXiv

  2. Vasilis Kontonis, Fotis Iliopoulos, Khoa Trinh, Cenk Baykal, Gaurav Menghani, Erik Vee

    SLaM: Student-Label Mixing for Distillation with Unlabeled Examples

    NeurIPS 2023
    Read on arXiv

  3. Cenk Baykal, Khoa Trinh, Fotis Iliopoulos, Gaurav Menghani, Erik Vee

    Robust Active Distillation

    ICLR 2023
    Read on arXiv

  4. Vasilis Kontonis, Fotis Iliopoulos, Cenk Baykal, Gaurav Menghani, Khoa Trinh, Erik Vee

    Weighted Distillation with Unlabeled Examples

    NeurIPS 2022
    Read on arXiv

Model Compression & Pruning

  1. Cenk Baykal*, Lucas Liebenwein*, Igor Gilitschenski, Dan Feldman, Daniela Rus

    SiPPing Neural Networks: Sensitivity-informed Provable Pruning of Neural Networks

    SIAM Journal on Mathematics of Data Science, 2022
    Read on arXiv

  2. Lucas Liebenwein, Cenk Baykal, Brandon Carter, David Gifford, Daniela Rus

    Lost in Pruning: The Effects of Pruning Neural Networks beyond Test Accuracy

    MLSys 2021
    Read on arXiv

  3. Lucas Liebenwein*, Cenk Baykal*, Harry Lang, Dan Feldman, Daniela Rus

    Provable Filter Pruning for Efficient Neural Networks

    ICLR 2020
    Read on arXiv

Theory & Foundations

  1. Cenk Baykal, Nishanth Dikkala, Rina Panigrahy, Cyrus Rashtchian, Xin Wang

    A Theoretical View on Sparsely Activated Networks

    NeurIPS 2022
    Read paper

  2. Cenk Baykal*, Murad Tukan*, Dan Feldman, Daniela Rus

    Coresets for Support Vector Machines

    Theoretical Computer Science, 2021 (extended version of TAMC 2020 oral)
    View publication

  3. Cenk Baykal*, Lucas Liebenwein*, Igor Gilitschenski, Dan Feldman, Daniela Rus

    Data-Dependent Coresets for Compressing Neural Networks with Applications to Generalization Bounds

    ICLR 2019
    Read on arXiv

Robotics

  1. Cenk Baykal*, Lucas Liebenwein*, Igor Gilitschenski, Sertac Karaman, Daniela Rus

    Sampling-Based Approximation Algorithms for Reachability Analysis with Provable Guarantees

    Robotics: Science and Systems (RSS), 2018
    Read PDF

  2. Cenk Baykal and Ron Alterovitz

    Asymptotically Optimal Design of Piecewise Cylindrical Robots using Motion Planning

    Robotics: Science and Systems (RSS), 2017 Best Paper Award
    Read PDF

Thesis

  1. Cenk Baykal

    Sampling-based Algorithms for Fast and Deployable AI

    Ph.D. Thesis, Massachusetts Institute of Technology, September 2021
    Read thesis

Curriculum Vitae

Download the latest CV

Download a concise summary of my current projects, publications, and experience. Last updated May 2026.

Download PDF

Connect

Reach out

Email works best for me. Feel free to reach out with questions or potential collaborations.

baykal@alum.mit.edu