PeopleofSuperintelligence

Researcher

Rafael Mosquera-Gomez

Also known as Rafael Mosquera, Rafael Mosquera-Gómez

Rafael Mosquera-Gomez is a machine learning engineer at MLCommons and Factored AI who works on open datasets and benchmarks for multilingual and human-aligned AI. He co-authored the DataPerf benchmark suite in 2022 and the Speech Wikimedia dataset in 2023. In 2024, he was an author of the PRISM Alignment Dataset, which gathered feedback from 1,500 participants in 75 countries and won the Datasets and Benchmarks Best Paper award at NeurIPS 2024. In 2025, he and Sarah Luger wrote MLCommons' announcement of the Unsupervised People's Speech dataset, and in January 2026 he co-chaired the LM4UC workshop at AAAI. He is a core contributor to the MLCommons Datasets Working Group and organizes the Interspeech 2026 Unsupervised Speech in the Wild Challenge.

Now at
MLCommons and Factored AI

Timeline

  1. 2022

    The DataPerf benchmarks

    In July 2022, he is one of the authors of DataPerf, a community-led benchmark suite for evaluating ML datasets and data-centric algorithms. Its first iteration holds five benchmarks in vision, speech, acquisition, debugging, and diffusion prompting.

  2. 2023

    Speech Wikimedia dataset

    In 2023, he is first author of Speech Wikimedia, a dataset of 1780 hours of CC-BY-SA licensed transcribed speech in 77 languages taken from Wikimedia Commons. The paper is a poster at the Data-centric Machine Learning Research workshop at ICML 2023.

  3. 2024

    The PRISM Alignment Dataset

    In 2024, he is an author of the PRISM Alignment Dataset, which maps the sociodemographics and stated preferences of 1,500 participants from 75 countries to their feedback in 8,011 live conversations with 21 language models. As an MLCommons engineer and author, he works on a project that extends the Dynabench platform with LLM plugins, new rating scales, and a chat interface.

  4. 2024

    A NeurIPS best paper award

    The PRISM paper, with twelve authors including him, wins the Datasets and Benchmarks Best Paper award at NeurIPS 2024. Hannah Rose Kirk and Juan Ciro are among his co-authors.

  5. 2025

    Unsupervised People's Speech

    In January 2025, he and Sarah Luger write MLCommons' announcement of the Unsupervised People's Speech dataset, built with Hugging Face and holding over 1 million hours of audio across dozens of languages. The working group's pipelines find 821,412+ hours of speech in 89 languages. A 2025 LoResMT paper he co-authors for the MLCommons Datasets Working Group describes the dataset and a Common Crawl collaboration on low-resource languages.

  6. January 2026

    Co-chairing LM4UC at AAAI

    In January 2026, he is one of eight co-chairs of the Language Models for Underserved Communities (LM4UC) workshop at the AAAI Conference in Singapore, alongside Sarah Luger and Sanmi Koyejo.

  7. 2026

    The CommonLID benchmark

    In 2026, he co-authors CommonLID, a community-driven, human-annotated language identification benchmark for web text that covers 109 languages. The paper lists him with Factored AI and MLCommons, includes co-authors from Common Crawl Foundation and EleutherAI, and appears in the proceedings of the 64th Annual Meeting of the Association for Computational Linguistics.

  8. As of October 2026

    Engineer at MLCommons, Factored AI

    He is a machine learning engineer at MLCommons and Factored AI and a core contributor to the MLCommons Datasets Working Group. He is an organizer of the Unsupervised Speech in the Wild Challenge at Interspeech 2026, which opens in December 2025 and trains self-supervised speech models on the Unsupervised People's Speech dataset.

Career

  1. NowMLCommonsMachine learning engineer
  2. NowFactored AIMachine learning engineer

From the biography.

  1. 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.
  2. 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.
  3. 43% of living people shaping AI work outside their country of birthWe estimate that 43% of living people shaping AI work outside their country of birth (520 of 1,198 people with known birth and employment countries).
  4. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionWe estimate where the people shaping AI studied and where their careers led. Massachusetts Institute of Technology leads with 467 former students, including 291 recorded doctorates.

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 Rafael Mosquera-Gomez yet. Send us a free-licensed photo.

  1. DataPerf: Benchmarks for Data-Centric AI Development arxiv.org. Checked 22 October 2026.
  2. Speech Wikimedia: A 77 Language Multilingual Speech Dataset | ICML 2023 icml.cc. Checked 22 October 2026.
  3. Speech Wikimedia: A 77 Language Multilingual Speech Dataset arxiv.org. Checked 22 October 2026.
  4. The PRISM Alignment Dataset: What Participatory, Representative and Individualised Human Feedback Reveals About the Subjective and Multicultural Alignment of Large Language Models arxiv.org. Checked 22 October 2026.
  5. Unveiling the PRISM Alignment Project - MLCommons mlcommons.org. Checked 22 October 2026.
  6. NeurIPS 2024 Awards neurips.cc. Checked 22 October 2026.
  7. Congratulations to the #NeurIPS2024 award winners - AIhub aihub.org. Checked 22 October 2026.
  8. Unsupervised People's Speech: A Massive Multilingual Audio Dataset - MLCommons mlcommons.org. Checked 22 October 2026.
  9. Building Data Infrastructure for Low-Resource Languages - ACL Anthology aclanthology.org. Checked 22 October 2026.
  10. Reports of the Workshops Held at the 2026 AAAI Conference on Artificial Intelligence | Interactive AI Magazine interactiveaimag.org. Checked 22 October 2026.
  11. AAAI 2026 Workshop: Language Models for Underserved Communities lm4uc.github.io. Checked 22 October 2026.
  12. CommonLID: Re-evaluating State-of-the-Art Language Identification Performance on Web Data arxiv.org. Checked 22 October 2026.
  13. Rafael Mosquera - ACL Anthology aclanthology.org. Checked 22 October 2026.
  14. Unsupervised Speech in the Wild Challenge | Interspeech 2026 upschallenge.org. Checked 22 October 2026.