Staff AI Engineer

📋 Description

• • Architect the next generation of intelligent strength training. You will own the end-to-end lifecycle of machine-learning systems that turn Tonal’s multi-modal dataset—millions of workouts enriched by cable sensors, 3-D cameras, and external trackers—into real-time, personalized coaching experiences for every member.
• • Design and optimize high-throughput training pipelines that can re-train transformer-based and self-supervised models on petabytes of time-series biomechanics data without impacting production latency. You will select the right mix of distributed training, mixed-precision optimization, and model-compression techniques so that new intelligence reaches members within hours, not weeks.
• • Build adaptive algorithms that go beyond “add five pounds next week.” Your models will reason over force curves, range-of-motion signatures, velocity decay, and perceived exertion to prescribe micro-adjustments in weight, tempo, and rest that continuously maximize strength gains while minimizing injury risk.
• • Ship production-grade Python services that serve sub-100 ms predictions at global scale. You will integrate tightly with GoLang and Kotlin backend teams, containerize models, define gRPC contracts, and implement canary releases so that every new insight rolls out safely to hundreds of thousands of homes.
• • Pioneer on-device inference. You will compress and quantize models so that critical coaching cues—like “brace your core” or “slow the eccentric”—can be delivered even when Wi-Fi drops, ensuring an uninterrupted member experience.
• • Create automated experimentation and validation frameworks. From offline A/B simulators to live shadow deployments, you will give data scientists and exercise physiologists self-service tools that accelerate the discovery of what actually drives progress.
• • Translate exercise-science research into code. You will partner with PhDs in biomechanics to encode the latest findings on fatigue, hypertrophy, and motor learning into differentiable loss functions and reward models.
• • Mentor a growing AI/ML guild. Through design reviews, brown-bag sessions, and pair programming, you will raise the bar on code quality, MLOps hygiene, and ethical AI practices across the company.
• • Influence product roadmap. By surfacing unexpected patterns in the data—like how hip-dominant athletes respond differently to tempo changes—you will inspire entirely new features that redefine what a home gym can do.
• • Champion inclusive technology. You will ensure that models perform equitably across age, gender, body type, and ability by curating balanced datasets, auditing outcomes, and iterating on fairness constraints.