SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Generalist is building embodied foundation models focused on dexterity, enabling robots to intelligently interact with the physical world. The company combines large-scale AI and robotics, with a team of researchers and roboticists from OpenAI, Boston Dynamics, Google DeepMind, and other frontier labs.
As an Applied Scientist, you will own the mastery of new frontier model capabilities by developing post-training recipes for learning dexterous tasks across diverse applications—from science lab automation to manufacturing and warehouse logistics. You will define evaluation criteria, build measurement protocols, run rigorous experiments across model variants and data recipes, and drive the post-training loop to compound capabilities that generalize across industries.
Key responsibilities include:
- Owning end-to-end improvement of target capabilities from baseline measurement to demonstrated mastery
- Designing task suites and evaluation protocols that capture research, science partner, and customer needs while generalizing beyond single deployments
- Coordinating data collection, training, fine-tuning, sweeping data mixtures and configurations, and orchestrating large-scale evaluations
- Running rigorous experimental analysis across model variants—isolating variables, quantifying statistical confidence, and diagnosing failure modes from rollout data
This is an empirical, model science-first role requiring excellence in research, engineering, and operations. You will leverage post-training, data, task design, and rigorous experimentation to build capabilities that become strengths of the foundation model.
REQUIREMENTS:
- Strong background in machine learning and/or robot learning (e.g., PhD in a relevant area)
- Comfort with data analysis at scale: slicing rollout logs, computing metrics, visualizing results, and investigating outcomes
- Deep care for evaluation, benchmarking, and failure analysis with diligence in face of repetitive but subtly changing task evaluations
- Ability to iterate rapidly with real-world feedback loops and willingness to work hands-on with robots and physical task setups
- Motivation to make capabilities work for real applications
- Strong communication skills for both research teams and partner-facing teams
