PeopleofSuperintelligence

Researcher

Jan Ludziejewski

Jan Ludziejewski is a machine-learning researcher whose work spans computer-vision tracking, mixture-of-experts language models and reasoning systems. He worked at QED Software and IDEAS NCBR while pursuing doctoral research at the University of Warsaw. In 2024, he coauthored ICML and NeurIPS studies of MoE scaling and continuous token routing. His 2025 work includes a study of memory-efficient MoE models and Magistral, Mistral AI's first reasoning model. He earned a University of Warsaw doctorate in 2026 and is an AI Research Scientist at Mistral AI, where he coauthored 2026 papers on Ministral 3 and Voxtral TTS.

Timeline

  1. 2017–2025

    From QED Software to IDEAS

    He joins QED Software as a machine-learning specialist in 2017 and moves to IDEAS NCBR as a doctoral researcher in 2022. In 2023, he is a doctoral student at the University of Warsaw's Institute of Computer Science and an IDEAS NCBR researcher, with interests in neural-network music generation.

  2. 2020

    Arahub human tracking

    In 2020, he coauthors a paper introducing Arahub, a system that combines computer-vision tracking with smartphone signals for interactive applications.

  3. 2024

    Fine-grained MoE scaling

    In 2024, he coauthors an ICML study that introduces granularity as a way to control expert size and derive mixture-of-experts scaling laws under a compute budget.

  4. 2024

    Mixture of Tokens

    In 2024, he coauthors Mixture of Tokens, a continuous architecture that assigns tokens from different examples to each expert and scales toward sparse MoE parameter counts.

  5. February 2025

    Memory-efficient MoE scaling

    In 2025, he coauthors Joint MoE Scaling Laws, which shows that MoE models can be more memory-efficient than dense models under fixed compute and memory budgets.

  6. June 2025

    Magistral reasoning

    In June 2025, he coauthors Magistral, Mistral AI's first reasoning model.

  7. June 2026

    A University of Warsaw doctorate

    In June 2026, he earns a computer-science doctorate at the University of Warsaw under supervisor Marek Cygan, with a dissertation on improving mixture-of-experts language models.

  8. Since March 2025

    Research scientist at Mistral

    Since March 2025, he is an AI Research Scientist in Mistral AI's Research department. In 2026, he coauthors Ministral 3, a family of parameter-efficient language models, and Voxtral TTS, an expressive multilingual text-to-speech model.

Career

  1. Since 2025Mistral AIAI Research Scientist
  2. 2022–2025IDEAS NCBRDoctoral researcher
  3. 2017–2022QED SoftwareMachine-learning specialist

Education

  1. From 2023University of WarsawPhD

From the biography.

  1. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Mistral AI. The median h-index is 48 at universities and 14 at frontier AI labs. We compare publishing careers, recent citations, dual appointments and job moves across 2,420 university researchers and 308 frontier-lab researchers.
  2. Companies supply the most prior-work hires at 7 of the 9 frontier AI labsIncludes Mistral AI. Our estimates of the organizations where 550 people worked directly before joining 9 frontier AI labs across 620 career transitions, comparing the main feeder organizations per lab.
  3. Of 943 people shaping AI who worked at Google, 446 work there nowIncludes Mistral AI. Our estimates of where the 943 people shaping AI who worked at Google and Google DeepMind work now: 446 at Google and 452 across 278 other organizations.
  4. 5,000 frontier users would be 38.2% of the core labs' 13,100 reported staffIncludes Mistral AI. 5,000 frontier users would be 38.2% of the core labs' 13,100 reported staff. We test the estimate against 3,913 named model contributors and dated all-staff records across Google DeepMind, OpenAI and Anthropic, plus six additional frontier-grade labs.

Colleagues

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

Sources and credits

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

  1. 10 Mistral AI research scientists instant.expert. Checked 20 October 2026.
  2. W rytmie AI. Jak wykorzystać sztuczną inteligencję w tworzeniu muzyki? Nowy odcinek podcastu ideas-ncbr.pl. Checked 20 October 2026.
  3. Integrated Human Tracking Based on Video and Smartphone Signal Processing within the Arahub System annals-csis.org. Checked 20 October 2026.
  4. Scaling Laws for Fine-Grained Mixture of Experts proceedings.mlr.press. Checked 20 October 2026.
  5. Mixture of Tokens: Continuous MoE through Cross-Example Aggregation proceedings.neurips.cc. Checked 20 October 2026.
  6. Joint MoE Scaling Laws: Mixture of Experts Can Be Memory Efficient arxiv.org. Checked 20 October 2026.
  7. Magistral arxiv.org. Checked 20 October 2026.
  8. Jan Ludziejewski mimuw.edu.pl. Checked 20 October 2026.
  9. Ministral 3 arxiv.org. Checked 20 October 2026.
  10. Voxtral TTS arxiv.org. Checked 20 October 2026.