PeopleofSuperintelligence

Researcher

Nouamane Tazi

Nouamane Tazi is a machine learning research engineer at Hugging Face who specializes in training large language models. He coauthored the MTEB text-embedding benchmark in October 2022 and was among the authors of the 176-billion-parameter BLOOM model in November 2022. Scaling Data-Constrained Language Models, which he wrote with eight co-authors, was named an Outstanding Main Track Runner-up at NeurIPS 2023. His work on open models since then includes StarCoder2 in February 2024, The Ultra-Scale Playbook on training across GPU clusters in July 2025, and the SmolLM3 model in July 2025.

Timeline

  1. October 2022

    The MTEB embedding benchmark

    In October 2022, he coauthors MTEB, the Massive Text Embedding Benchmark, with Niklas Muennighoff, Loïc Magne and Nils Reimers. The benchmark spans 8 embedding tasks, 58 datasets and 112 languages.

  2. November 2022

    Authoring the BLOOM model

    In November 2022, he is among the authors of BLOOM, a 176B-parameter open-access language model that the BigScience Workshop builds through a collaboration of hundreds of researchers. It is trained on a corpus covering 46 natural and 13 programming languages.

  3. 2023

    NeurIPS 2023 runner-up paper

    He is one of nine authors of Scaling Data-Constrained Language Models, posted in May 2023, which runs experiments up to 900 billion training tokens and 9 billion parameters to study scaling when text data is limited. NeurIPS 2023 names the paper an Outstanding Main Track Runner-up.

  4. February 2024

    StarCoder2 and The Stack v2

    In February 2024, he is a named author of StarCoder2, the BigCode project's family of open code language models with 3B, 7B and 15B parameters, built on The Stack v2. Hugging Face trains the 7B model.

  5. 2025

    The Ultra-Scale Playbook

    In April 2025, he gives The Ultra-Scale Talk on scaling training to thousands of GPUs at the Technology Innovation Institute's AI seminar series, covering 5D parallelism across data, tensor, pipeline, context and expert dimensions. He is first of seven authors of The Ultra-Scale Playbook, a Hugging Face book on training models at the largest scales, published in July 2025.

  6. July 2025

    The SmolLM3 model

    In July 2025, he is among the authors of SmolLM3, a fully open 3B-parameter model trained on 11 trillion tokens. It outperforms Llama-3.2-3B and Qwen2.5-3B and stays competitive with larger 4B models.

  7. As of October 2026

    Research engineer at Hugging Face

    He is a Machine Learning Research Engineer at Hugging Face, specializing in training large language models. In May 2026, he presents The Smol Training Handbook: Lessons from Training State-of-the-Art LLMs at GOSIM Paris 2026.

  1. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes Hugging Face. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.
  2. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes Hugging Face. 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.
  3. Google has the highest combined h-index among AI organizationsIncludes Hugging Face. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.
  4. Companies supply the most prior-work hires at 7 of the 9 frontier AI labsIncludes Hugging Face. 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.

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 Nouamane Tazi yet. Send us a free-licensed photo.

  1. MTEB: Massive Text Embedding Benchmark arxiv.org. Checked 22 October 2026.
  2. BLOOM: A 176B-Parameter Open-Access Multilingual Language Model arxiv.org. Checked 22 October 2026.
  3. Scaling Data-Constrained Language Models arxiv.org. Checked 22 October 2026.
  4. Announcing the NeurIPS 2023 Paper Awards blog.neurips.cc. Checked 22 October 2026.
  5. StarCoder2 and The Stack v2 huggingface.co. Checked 22 October 2026.
  6. StarCoder 2 and The Stack v2: The Next Generation arxiv.org. Checked 22 October 2026.
  7. AI Seminar Series: Nouamane Tazi tii.ae. Checked 22 October 2026.
  8. nanotron (Nanotron Research) huggingface.co. Checked 22 October 2026.
  9. SmolLM3: smol, multilingual, long-context reasoner huggingface.co. Checked 22 October 2026.
  10. Nouamane Tazi – GOSIM Paris 2026 paris2026.gosim.org. Checked 22 October 2026.