Research Engineer - Data Eval
Core
Designing and executing evaluation research for embodied AI and world models, including running ablations, model-based validation, and inventing benchmarks to quantify data quality and model performance.
Role type
Senior IC research engineer (data evaluation)
Builds
Evaluation pipelines, dashboards, and tooling for internal teams and external partners
Domain
Embodied AI, robotics, world models, computer vision
Deliverable
production ML models
Required skills
ML experimentation, experimental design, Python, scalable data pipelines, robotics, imitation learning, reward modeling, calibration and uncertainty quantification
Preferred skills
Large-scale data pipelines, vision-language-action models, multimodal egocentric perception
Technologies
PyTorch, open-source model ecosystems, simulators, real-world sensor stacks
Responsibilities
Running ablations to quantify collection and labeling choices; validating raw data for defects like motion blur or sensor desync; setting the evaluation research agenda; inventing benchmarks and metrics; running end-to-end model evaluation; building automated evaluation tooling; mentoring cross-functional teams
Seniority
Senior, hands-on IC with mentorship
