Focus: Build and ship the machine learning pipelines that give our devices real-time perception and intelligence, from prototype to production at scale. 3–6 years of experience working with AI models in production, ideally for real-time applications. Fluent in Python for data pipelines and backend features involving third-party model providers’ APIs and SDKs. Well-versed in a variety of data stores and their tradeoffs: relational DBs, No-SQL, cloud object stores, etc. Experience with modern ML frameworks (PyTorch/TensorFlow), and experience optimizing models for edge/embedded inference is a plus, but not required. Build data pipelines and evaluation frameworks to continuously improve AI feature quality from real-world usage. Own the full lifecycle: data collection, dataset versioning, evaluation design, deployment and monitoring in the field, and feature improvement recommendations. Comfortable with cloud infrastructure. Ability to move fast in an environment where things are sometimes loosely defined, and a genuine curiosity about where AI meets physical products.