Machine Learning Engineer - Representation Learning
Core
Design, train, and scale self-supervised pretraining methods for multimodal time-series data to build general-purpose models for real-world sensing applications.
Role type
Senior IC machine learning engineer (representation learning)
Builds
General-purpose models for wearable sensing and human behavior understanding
Domain
AI / Machine Learning / Time-series / Wearable Sensing
Deliverable
production ML models
Required skills
Deep learning, Self-supervised learning, Representation learning, Foundation model pretraining, PyTorch, Transformers, State-space models, Temporal CNNs, Distributed training, Mixed precision, Large-dataset pipelines, Model scaling, Fine-tuning, Adaptation methods, Evaluation protocols, Linear probing, Few-shot transfer
Preferred skills
Wearable sensing, Multimodal learning, Cross-modal alignment, Publications at NeurIPS/ICML/ICLR/CVPR/IMWUT/UbiComp
Technologies
PyTorch, Transformers, State-space models, Temporal CNNs
Responsibilities
Design and implement masked modeling, contrastive learning, and predictive/JEPA-style approaches for time-series data. Develop models that generalize across users, devices, and real-world conditions. Create fine-tuning and adaptation methods for downstream tasks. Establish evaluation protocols and benchmarks. Collaborate with data teams on corpus strategy. Work with hardware and sensing teams to incorporate modeling insights into sensor design. Help develop training infrastructure and experiment tracking.
Seniority
Senior, hands-on IC
