Jagiellonian University
Jonas RothfussMember of Technical Staff · Microsoft AIResearcher↗
Researcher
Konrad Żołna
Microsoft AIMember of Technical Staff
Żołna develops learning systems that combine large pretrained models with agents and robots. At Google DeepMind he co-created GATS, a way to join foundation models across modalities without erasing their learned knowledge[1], and advanced offline reinforcement learning from demonstrations and unlabeled experience.
Career & education
Current roleMicrosoft AI2024-presentMember of Technical Staff
Microsoft AI2024-presentPreviously
Google DeepMind2019-2024Research Scientist -> Senior Research Scientist
Google DeepMind2019-2024Education
Work and education history is primarily focused on AI-relevant roles and may not be comprehensive.
Citation Trend
Total citations to scholarly work, including research papers, technical reports, books and theses, as recorded by Google Scholar. Each point is a dated snapshot.
citations5,949As of Nov 7, '26
citations+1,077Since last snapshot, Jun 15, '26
Selected contributions
A Generalist Agent
With Scott Reed, Emilio Parisotto, Sergio Gomez Colmenarejo, Alexander Novikov, Gabriel Barth-Maron, Mai Gimenez, Yury Sulsky and others ·Gato brought language, games, image descriptions and robot control into a single model.
Genie: Generative Interactive Environments
With Jake Bruce, Michael D Dennis, Ashley Edwards, Jack Parker-Holder, Yuge Shi, Edward Hughes, Matthew Lai and others ·Genie learns from videos to turn an image into a controllable virtual world.
References
- GATS: Gather-Attend-Scatterarxiv.org
Co-authorship of GATS; combining pretrained models across modalities while keeping components frozen to preserve pretrained knowledge.
Based on public sources. If accuracy is important, verify details using the links or references.
For corrections, updates, or requests to remove information, contact us.
Reviewed August 9, 2026 · Checked for accuracy and new information.
