Machine Learning Engineer for Digital Twin

We're building a Digital Twin platform for physical therapists and fitness professionals. The role involves developing machine learning models for IMU/PPG signal processing and motion intent prediction. You'll work with a team to integrate these models into a cloud-based platform, enhancing the accuracy of human movement analysis. Experience with wearable devices and healthcare applications is a plus. Core tasks include building an SVM-based classifier to distinguish known from unknown movement patterns, and using LASSO regression alongside neural network models (LSTM/CNN) to predict motion intent. This is our current design direction — if you have a stronger approach based on your experience, we're open to alternatives, provided the solution still meets these requirements: (1) some level of interpretability suitable for clinical/coaching use, (2) feasible for real-time or near-real-time inference on wearable devices, and (3) able to handle open-set/novelty detection for movement patterns not seen during training. When applying, feel free to attach a simple architecture diagram alongside your proposal — this helps us understand your technical approach faster.