Biosignal/ML Engineer — Detect & Classify Jaw Activity from Wearable Sensor Data

We're an early-stage health-wearable startup building a device that distinguishes teeth grinding from clenching from ordinary jaw movement using motion, mechanical, and heart-rate sensors. We need a signal-processing/ML engineer to develop the detection logic from recorded sensor data. Scope of this engagement: Work with recorded multi-sensor data (motion, mechanical/muscle, optical) — initially analyzed on a computer, not yet on-device Build signal-processing and classification logic to: detect jaw-muscle activity, separate sustained clenching from rhythmic grinding, flag ordinary chewing/talking, and estimate intensity Validate the output against a reference signal (medical EMG) recorded alongside our sensors Document an approach that could later be miniaturized to run on the device (embedded/TinyML) — a plus, not required for this phase You're a great fit if you have: biosignal or time-series experience (EMG, ECG/PPG, IMU/activity recognition, accelerometer-based detection), solid signal processing and feature engineering, and practical ML for classification. Embedded/TinyML experience is a bonus. Please include: examples of biosignal or sensor time-series projects, and a short note on how you'd tell a sustained muscle contraction apart from a rhythmic one. We're starting with a defined analysis milestone with strong potential to grow into the core algorithm role.