PeopleofSuperintelligence

Researcher

Karina Zadorozhny

Karina Zadorozhny is a Principal Machine Learning Engineer at Prescient Design, Genentech, and a coauthor of work on generative models for antibodies and other biomolecules. They began in DNA replication biology, coauthoring a 2018 Nature paper on SAMHD1 from Masaryk University's Faculty of Medicine, and later wrote guidelines for out-of-distribution detection in medical machine learning with colleagues at Pacmed and Amsterdam UMC. At Prescient Design, they coauthored "Protein Discovery with Discrete Walk-Jump Sampling," an ICLR 2024 Outstanding Paper, and were first author of the 2025 SQRL paper on molecular activity prediction. They also appear among the authors of the 2025 lab-in-the-loop antibody design preprint and the 2026 LeFlur biomolecular design model.

Now at
Roche

Timeline

  1. 2017–2018

    DNA replication papers at Masaryk

    Zadorozhny is first author of the 2017 Cell Reports paper "Fanconi-anemia-associated mutations destabilize RAD51 filaments and impair replication fork protection." A 2018 Nature paper from Masaryk University's Faculty of Medicine, coauthored by Zadorozhny, shows that SAMHD1 promotes degradation of nascent DNA at stalled replication forks.

  2. 2021–2022

    Medical out-of-distribution detection guidelines

    From Pacmed in Amsterdam and the Laboratory for Critical Care Computational Intelligence at Amsterdam UMC, Zadorozhny writes with Patrick Thoral, Paul Elbers and Giovanni Cinà. Their paper "Out-of-Distribution Detection for Medical Applications: Guidelines for Practical Evaluation," submitted in September 2021 and revised in May 2022, proposes practical tests for choosing an out-of-distribution detector for a medical dataset and illustrates them on electronic health records.

  3. June 2023

    Walk-jump sampling for antibodies

    The paper "Protein Discovery with Discrete Walk-Jump Sampling," posted in June 2023, has Zadorozhny of Prescient Design, Genentech among its authors, with Nathan Frey, Kyunghyun Cho, Vladimir Gligorijevic and others. On antibody proteins, 97–100% of generated samples are expressed and purified, and 70% of functional designs match or improve the binding affinity of known antibodies in a single round of lab experiments.

  4. 2024

    An ICLR Outstanding Paper award

    The walk-jump sampling paper wins an Outstanding Paper award at ICLR 2024 and is presented there as an oral.

  5. January 2025

    SQRL for molecular activity prediction

    Zadorozhny is first author of "Similarity-Quantized Relative Difference Learning for Improved Molecular Activity Prediction," submitted in January 2025 with Kangway Chuang, Bharath Sathappan, Ewan Wallace, Vishnu Sresht and Colin Grambow. The SQRL framework treats molecular activity prediction as relative difference learning between structurally similar compound pairs and improves accuracy in the low-data regimes common in drug discovery.

  6. February 2025

    Lab-in-the-loop antibody design

    In February 2025, the bioRxiv preprint "Lab-in-the-loop therapeutic antibody design with deep learning" appears with Zadorozhny among its authors, alongside Nathan Frey, Richard Bonneau, Kyunghyun Cho and Vladimir Gligorijevic. It links generative models, property predictors, active learning and lab experiments in an iterative loop, designs and tests over 1,800 antibody variants across four targets, and reports 3-100x better binders for every target after four rounds.

  7. 2026

    The LeFlur design model

    A paper by Zadorozhny, Sidney Lisanza, Frederic Dreyer and Kyunghyun Cho, "LeFlur: A Biomolecular Design Model with Latent Structure Tokens," is part of the ICML 2026 workshop Generative and Agentic AI for Biology.

  8. As of October 2026

    Principal ML engineer at Genentech

    Zadorozhny is a Principal Machine Learning Engineer at Prescient Design, Genentech.

  1. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes Roche. We estimate where the people shaping AI worked when their most-cited papers appeared: the company share was 31% in 2020–2024 and 8% in 1970–1974, and Google led all organizations in 2020–2024.
  2. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes Roche. We estimate that 539 of the 6,766 people with a known current organization (8%) hold both a university role and a company role, including 20% of people at companies and 13% of people at universities.
  3. Google had the most people shaping AI every year from 2014 to 2025Our estimates of which organizations employed the most people shaping AI each year from 1955 to 2025, tracing the shift from university labs to companies.
  4. The United States employs an estimated 48% of people shaping AIWe estimate where the people shaping AI work today, how universities, companies and research institutes split across countries, and how country shares have changed since the 1960s.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

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

  1. SAMHD1 acts at stalled replication forks to prevent interferon induction | Masaryk University | MUNI muni.cz. Checked 21 October 2026.
  2. Karina Zadorozhny - Google Scholar scholar.google.com. Checked 21 October 2026.
  3. [2109.14885] Out-of-Distribution Detection for Medical Applications: Guidelines for Practical Evaluation arxiv.org. Checked 21 October 2026.
  4. Out-of-Distribution Detection for Medical Applications: Guidelines for Practical Evaluation arxiv.org. Checked 21 October 2026.
  5. [2306.12360] Protein Discovery with Discrete Walk-Jump Sampling arxiv.org. Checked 21 October 2026.
  6. Protein Discovery with Discrete Walk-Jump Sampling arxiv.org. Checked 21 October 2026.
  7. ICLR 2024 Awards iclr.cc. Checked 21 October 2026.
  8. [2501.09103] Similarity-Quantized Relative Difference Learning for Improved Molecular Activity Prediction arxiv.org. Checked 21 October 2026.
  9. Lab-in-the-loop therapeutic antibody design with deep learning twistbioscience.com. Checked 21 October 2026.
  10. Generative and Agentic AI for Biology icml.cc. Checked 21 October 2026.