Data Scientist — Wearable & Biometric Analytics

Marco Russo

Contract-to-hire · Mid-level · Los_Angeles

This role is no longer accepting applications

Similar open roles you can apply to now:

About the role

We make a recovery and sleep tracking ring. 2.8 million active users. 60 billion biometric readings in our data lake at last count — heart rate, HRV, skin temperature, SpO2, accelerometry, and a growing set of derived signals our algorithms compute from combinations of those inputs. Our users wear this device every night because they trust what it tells them about their body. That trust is the product. The data science team\'s job is to make sure the insights we surface are accurate, meaningful, and — ideally — things that our users couldn\'t have known without us. We are hiring a Data Scientist for a six-month contract with the expectation of conversion to permanent. The work is in our core analytics group, which owns the signal processing, health metric derivation, and personalisation models that power the app. Specifically, we need someone to work on sleep stage classification accuracy improvements and on the expansion of our stress and recovery score models to new user populations where our training data is currently underrepresented.

Responsibilities

Improve sleep stage classification model accuracy across diverse user populations, with particular focus on underrepresented demographics in the current training set
Develop and validate stress and recovery score models for user groups where our current models show performance degradation
Design evaluation frameworks that account for individual variability in physiological signals — population-level accuracy metrics are necessary but insufficient
Collaborate with algorithm engineers to translate validated models into the signal processing pipeline
Contribute to the team\'s data labelling and clinical validation study design

Requirements

4+ years of data science with experience in time-series analysis, physiological signal processing, or health analytics
Python at a professional level: clean, tested, production-ready analysis and modelling code
PyTorch for developing and fine-tuning deep learning models on multivariate physiological time-series data
NumPy and Scikit-learn for signal processing, feature engineering, and baseline model development
Statistics at depth: Bayesian methods, mixed-effects models, and the ability to reason carefully about model generalisation across population subgroups
SQL for cohort analysis, label extraction from clinical validation studies, and user behaviour analysis at scale
Familiarity with consumer health data: HIPAA/GDPR considerations, the distinction between consumer wellness claims and medical device claims, and the evaluation standards appropriate to each

Benefits

$112,000 – $138,000 annualised contract rate, converting to competitive permanent salary
6-month contract with strong intention to convert — we do not use contract-to-hire as a long probation, we do it because the scope is genuinely project-shaped at the start
Full remote — US and Canada preferred for time zone overlap
A device and subscription, obviously
Access to one of the largest consumer biometric datasets in the world — it is genuinely unusual data to work with