About Dreem
Each and every night, around the world, hundreds of millions of people are trying to sleep but can’t. We’re in the middle of a global sleep crisis with serious ramifications for society. At Dreem, a team of 50 experts is on a mission to beat insomnia, build new ways to analyze sleep and build the sleep medicine of tomorrow.
In the 7 years since Dreem’s creation, we’ve collected 2M nights of sleep data and helped thousands of people improve their sleep. We also accumulated 45 patents and completed 3 full hardware iterations and 100s of software iterations. We raised $60M from investors including Laurent Alexandre, Xavier Niel, MAIF, BPI and Johnson & Johnson - the world’s healthcare leader. Along the way, we created a research network of 200 leading laboratories and hospitals and a scientific board of 4 of the most renowned neuroscientists and sleep experts in the world.
The company has been building a digital sleep clinic, providing comprehensive patient care to answer the growing needs of people suffering from sleep disorders - from screening, diagnosis, to the treatment of insomnia and apnea. By facilitating access to remote care, while improving diagnosis and treatment methods, the Dreem clinic is a unique alliance of advanced technology and sleep medicine, providing US payers with lower costs, improved clinical performance, and increased patient satisfaction.
Dreem is a fast-paced and dynamic work environment. The team’s fields of expertise encompass sleep research, neuroscience, mechanical engineering, machine learning, electronic design, embedded systems, testing, ergonomics, development of proprietary applications and software, design, marketing, and much more. Joining the team is challenging but hugely rewarding- working at Dreem means pushing the boundaries and being able to bring real positive change to people around the world. Ready to join the fight for better sleep?
About the role
Importance of the role (a.k.a. this is why we need you)
At Dreem, with more than two million nights analyzed (with deep learning !), data and algorithms are in our DNA. Our clients need to have access to robust and accurate analyses of their nights to diagnose, understand, and solve sleep issues that affect millions of people. In that role, you will work specifically on our sleep apnea detection models to improve their performances, computational efficiency, and robustness to allow their use in real life. You will also collaborate on our internal Deep Learning library to make your new models available to the team.
Job Description
- Work on a literature review to spot the latest methods that may be useful for Sleep Apnea detection
- Improve and optimize our Sleep Apnea detection models to reach better performances or computational efficiency on our internal datasets from Dreem devices
- Process external and internal datasets to train and evaluate models
- Work with the Deep Learning Library to make it easier to use, more computationally efficient, or more robust
Key Missions
- Literature review on Deep Learning and more specifically on Deep Learning applied to events detection on multi-modal biosignal data
- Implementation and improvement of the latest deep learning models applied to sleep apnea detection
- Ensure that your models' implementations are easy to use, well-documented, and well-tested by following best practices (Unit test, documentation, CI, non-regression, …) throughout all your developments
Approximate timeline
-
Within 1 month: You will have performed a literature review and models from our internal library on available datasets
-
Within 3 months: You will have implemented recent models from the literature and compared them with our internal models
-
Within 6 months: You will have proposed, implemented, and evaluated possible improvements of the model from the literature to fit them to our needs (Optimization to run on a smartphone, performances on multimodal data, robustness to noise)
About you (no need to tick all the boxes!):
- Last year student in an MSc/PhD in Computer Science, Machine learning, or a quantitative field
- Strong skills in Deep Learning, proven via successful experience(s)/project(s)
- Best practices in software engineering (test, Github, documentation, Dockerization, CI/CD) or willingness to learn and apply them
- Experience with Torch, Keras or Tensorflow to train, develop and deploy Deep learning model
- Rigor and willingness to learn new topics
- Fluent in English
Additional Information
🏡 Great office in the center of Paris (75003)
🍽 Lunchr (Tickets Restaurants)
🏃♂️ Gymlib
🚈 50% pass Navigo or bike indemnities
😴 Sleep with a headband
