Raphael Vallat, PhD

About

I am a Staff Machine Learning Scientist at Oura, where I build and ship health-tracking algorithms from concept to production. I lead end-to-end ML development on large wearable datasets, combining physiological insight with modern modeling approaches to build foundation models and digital health screening algorithms across both wellness and regulated contexts.

A core part of my work is translating deep learning research into systems that operate at scale and have real health impact. At Oura, I act as a technical lead on cross-functional teams, collaborating closely with product, clinical, and regulatory partners to bring deep learning models from research into production.

Trained as a neuroscientist, my background is in sleep physiology, neuro-imaging and wearable sensing. Before Oura, I was a postdoctoral researcher at UC Berkeley's Center for Human Sleep Science, studying how sleep disruption contributes to cardio-metabolic and neurodegenerative disease. My research has been published in top-tier journals and featured in major media outlets.

Alongside my industry and research work, I enjoy building tools for the broader scientific community. I am the creator and maintainer of widely used open-source Python libraries, including Pingouin (statistics), YASA (sleep analysis), and AntroPy (time-series complexity).

Publications

Please find below a list of selected publications (last updated May 2025)
PDF versions are provided for individual, noncommercial purposes only. These files may not be reposted without permission. Copyright and all rights therein resides with the respective copyright holders, as stated within each paper.

• Willoughby, Vallat, Ong, and Chee (2025). Insights about travel-related sleep disruption from 1.5 million nights of data. SLEEP.
• Vallat, Shah, and Walker (2023). Coordinated human sleeping brainwaves map peripheral body glucose homeostasis. Cell Reports Medicine.
• Vallat, Berry, ..., and Walker (2022). How people wake up is associated with previous night's sleep together with physical activity and food intake. Nature Communications.
• Ben Simon, Vallat, Rossi, and Walker (2022). Sleep loss leads to the withdrawal of human helping across individuals, groups, and large-scale societies. PLoS Biology.
• Tsereteli, Vallat, ..., and Franks (2021). Impact of insufficient sleep on dysregulated blood glucose control under standardised meal conditions. Diabetologia.
• Vallat and Walker (2021). An open-source, high-performance tool for automated sleep staging. eLife.
• Vallat, Nicolas and Ruby (2020). Brain functional connectivity upon awakening from sleep predicts inter-individual differences in dream recall frequency. SLEEP.
• Vallat*, Shah*, Redline, Attia and Walker (2020). Broken sleep predicts hardened blood vessels. PLoS Biology. * Co-first authors
• Ben Simon*, Vallat*, Barnes and Walker (2020). Sleep loss and the socio-emotional brain. Trends In Cognitive Sciences. * Co-first authors
• Vallat and Ruby (2019). Is it a good idea to cultivate lucid dreaming? Frontiers in Psychology.
• Plailly, Villalba, Vallat, Nicolas, and Ruby (2019). Incorporation of fragmented visuo-olfactory episodic memory into dreams and its association with memory performance. Scientific Reports, 9, 15687.
• Vallat (2018). Pingouin: statistics in Python. Journal of Open Source Software, 3(31), 1026.
• Vallat, Meunier, Nicolas, and Ruby (2018). Hard to wake up? The cerebral correlates of sleep inertia assessed using combined behavioral, EEG and fMRI measures. NeuroImage [PDF]
• Vallat, Eichenlaub, Nicolas, and Ruby (2018). Dream recall frequency is associated with medial prefrontal cortex white-matter density. Frontiers in Psychology, 9, 1856.
• Vallat (2017). Content and frequency of dream reports: psychological and neurophysiological correlates. PhD thesis, Lyon 1 University.
• Combrisson*, Vallat*, ..., and Ruby (2017). Sleep: an open-source python software for visualization, analysis and staging of sleep data. Frontiers in Neuroinformatics, 11. * Co-first authors
• Vallat, Chatard, Blagrove, and Ruby (2017). Characteristics of the memory sources of dreams: a new version of the content-matching paradigm to take mundane and remote memories into account Plos One, 12.

Media outreach

• New Study from NUS Analyzes Jet Lag and Sleep Patterns From 1.5 Million Nights of Oura Data @ ŌURA, The Pulse Blog
• Inside the Ring: Developing Oura’s Latest Sleep Staging Algorithm @ ŌURA, The Pulse Blog
• 5 ways to prepare for Daylight Saving Time now – and why your health depends on it @ Women's Health
• A Good Night’s Sleep May Help Control Blood Sugar @ Scientific American
• New research finds deep-sleep brain waves predict blood sugar control @ Berkeley News
• Sleepyheads need these three habits to get rid of grogginess @ New York Post
• Scientists discover secret to waking up alert and refreshed @ Berkeley News
• Sleep and metabolism @ The Gut Check Series
• The sad truth about sleep-tracking devices and apps @ New York Times
• Why do some people always remember their dreams, while others almost never do? @ Discover Magazine
• Vivid dreams and their role in waking life @ WHYY (NPR)
• The paradox of dreams @ This Is Your Brain
• Fitful nightly sleep linked to chronic inflammation, hardened arteries @ Berkeley News
• Simultaneous EEG and fMRI measure sleep inertia @ PhysicsWorld.com
• Brain scans reveal why it takes so long to wake up in the morning @ Inverse.com

Open-source

Please visit my GitHub repository for an exhaustive list of the projects/software that I am contributing to.

Pingouin

Pingouin is an open-source Python package for statistics. It is built on top of Pandas and NumPy and provides high-level functions to perform the most common statistical tests. Learn more on the documentation or check the code on GitHub

YASA

YASA (Yet Another Spindle Algorithm) is an open-source Python package for sleep analysis. It implements several state-of-the-art and validated algorithms to score sleep stages, detect spindles & slow-waves, and calculate features from sleep polysomnography recordings. Learn more on the documentation or check the code on GitHub

AntroPy

AntroPy is an open-source Python package for computing several entropy & complexity metrics of time-series. Learn more on the documentation or check the code on GitHub.