PeopleofSuperintelligence

Researcher

Rachid Riad

Also known as Rachid Adam Riad

Rachid Riad is a speech and machine-learning researcher whose work spans convolutional networks, clinical speech analysis and mental health care. Trained at Ponts Paristech and in the MVA master's program at ENS Paris-Saclay, he worked at ENS Paris under Emmanuel Dupoux and Anne-Catherine Bachoud-Lévi on algorithms that measure the speech of people with Huntington's disease. In 2022, his ICLR paper on DiffStride, a downsampling layer with learnable strides written with Olivier Teboul, David Grangier and Neil Zeghidour, received an Outstanding Paper award. He co-founded the speech-AI startup Callyope in 2023, published a 2024 Journal of Medical Internet Research study on speech and depression, and coauthored the SLAP audio-language model in 2025. He is a co-founder of Callyope, a French startup that monitors mental health symptoms from speech.

Now at
Callyope
Badges

Timeline

  1. 2018

    Siamese networks for speech

    In 2018, he coauthors a paper with Corentin Dancette, Julien Karadayi, Neil Zeghidour, Thomas Schatz and Emmanuel Dupoux on sampling strategies in Siamese networks for unsupervised speech representation learning. The work is presented at Interspeech 2018 in Hyderabad.

  2. 2020–2022

    Speech tests for Huntington's disease

    A graduate of Ponts Paristech and of the MVA master's program in computer vision and machine learning at ENS Paris-Saclay, he works between the LSCP and NPI labs at ENS Paris under Emmanuel Dupoux and Anne-Catherine Bachoud-Lévi. He builds algorithms that measure the speech of people with Huntington's disease and presents vocal markers from sustained phonation at Interspeech in 2020. In 2022, he is first author of a Journal of Neurology paper whose speech features predict cognitive, motor and functional scores with a relative error of 12.7 to 20.0%.

  3. 2022

    An ICLR 2022 outstanding paper

    With Olivier Teboul, David Grangier and Neil Zeghidour, he introduces DiffStride, a downsampling layer with learnable strides that learns the size of a cropping mask in the Fourier domain. The paper "Learning Strides in Convolutional Neural Networks" is an oral presentation at ICLR 2022 and receives an Outstanding Paper award.

  4. 2023

    Co-founding Callyope

    In 2023, he co-founds the French startup Callyope with Martin Denais, its CEO, and Xuan-Nga Cao. The company builds a speech foundation model that monitors symptoms of depression, anxiety, insomnia and psychosis from patient speech, smartphone sensors and clinical information.

  5. 2024

    Speech analysis for mental health

    In 2024, he is first author of a Journal of Medical Internet Research study at Callyope in which 865 healthy adults record answers about how they feel and whether they have slept well. A pipeline built on pretrained speech models detects and estimates depression, anxiety, insomnia and fatigue from the recordings, and the best model is Whisper M with max pooling and oversampling.

  6. 2025

    The SLAP audio-language model

    In 2025, he coauthors SLAP with Angelika Ando, Auguste Crabeil and Adrien Lesage at Callyope. The model aligns speech with natural language descriptions of speaker and health metadata through contrastive learning and reaches 62.9% average F1 in zero-shot evaluation, a 48% relative improvement over CLAP.

  7. As of October 2026

    Callyope co-founder

    He is a co-founder of Callyope, the French startup founded in 2023. It runs seven clinical studies with more than 1,000 patients, including a strategic R&D partnership with GHU Paris.

Career

  1. Since 2023CallyopeCo-founder
  2. 2020–2022École Normale SupérieureResearcher

Education

  1. École des Ponts ParisTech
  2. École Normale Supérieure Paris-SaclayMaster's degree

From the biography.

  1. The median h-index is 48 at universities and 14 at frontier AI labsIncludes École Normale Supérieure. 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. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes École Normale Supérieure. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.
  3. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes École Normale Supérieure. We 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.
  4. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes École Normale Supérieure. 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.

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 Rachid Riad yet. Send us a free-licensed photo.

  1. About Me - Rachid Riad rachine.github.io. Checked 22 October 2026.
  2. Predicting clinical scores in Huntington’s disease: a lightweight speech test pmc.ncbi.nlm.nih.gov. Checked 22 October 2026.
  3. ICLR Oral Learning Strides in Convolutional Neural Networks iclr.cc. Checked 22 October 2026.
  4. Callyope - LLM revolutionising psychiatric monitoring / AI World aiworld.eu. Checked 22 October 2026.
  5. Callyope - Google for Startups startup.google.com. Checked 22 October 2026.
  6. Automated Speech Analysis for Risk Detection of Depression, Anxiety, Insomnia, and Fatigue: Algorithm Development and Validation Study ncbi.nlm.nih.gov. Checked 22 October 2026.
  7. SLAP: Learning Speaker and Health-Related Representations from Natural Language Supervision arxiv.org. Checked 22 October 2026.