Robotics Research Engineer at Datoric

As a founding Robotics Research Engineer, you will lead the data design loop, identifying why manipulation policies fail and designing controlled experiments to fix them. You’ll use proxy models and weight space programming with SOTA LLMs to validate data hypotheses, eventually scaling successful collection processes into high-quality, real-world datasets for physical AI.

• Analyze manipulation policies to identify specific failure modes in generalization and recovery, then design controlled comparisons to test data-driven hypotheses.
• Train and deploy proxy models to screen multiple data collection strategies in parallel, ensuring high-confidence feedback before large-scale lab implementation.
• Develop repeatable data collection pipelines that manage hardware, sensors, and human contributors to produce high-quality datasets at industrial scale.

• Extensive experience training and debugging robotic manipulation policies, with a deep understanding of why models fail in real-world environments.
• Strong background in experimental design and evaluation, capable of designing ablations that settle complex questions about model behavior and data quality.
• Technical generalist with a high innovation mindset, comfortable moving between research, hardware, and operations to build physical AI infrastructure.

• Unprecedented compensation: $500k to $800k total compensation with significant founding-team equity at a rapidly growing YC S26 startup.
• Cutting-edge research: Work at the frontier of physical AI, utilizing weight space programming to train VLAs and debugging SOTA manipulation policies.
• Founding impact: Influence the foundational data strategy for the entire robotics industry, moving beyond “more demonstrations” to structured, scientific data design.