Human Data Manager (Robotics)

Job Title: Human Data Manager (Robotics) Job Type: Full-time Location: Remote About Us: The AI platform for human intelligence—we build models that rigorously vet PhDs and scale human data training for frontier LLMs. Our agent, Zara, autonomously sources and evaluates researchers and data trainers—powering post-training (RLHF, evals, red-teaming) for top AI labs. We've already built the infra layer that teaches language models how to think. Now, we're using that same layer to teach robots how to act. About the Role Robotics faces a major data bottleneck: capturing real-world, high-quality demonstrations—as a Human Data Manager, you'll run our human data engine — turning raw human and robot interactions into structured, training-ready data for large-scale robotics and VLA models. What You'll Do Manage large-scale workflows for human demonstration data collection (home, office, tools, etc.) using cameras, wearables, and sensors. Oversee robot teleoperation data capture — ensuring synchronized video, control signals, and quality consistency. Build and optimize annotation and QA pipelines for accuracy, structure, and scalability. Track operational KPIs and analyze data to improve efficiency and model-readiness. Collaborate with engineering, product, and QA to align on dataset standards and delivery goals. What We're Looking For Bachelor's or Master's in Computer Science, Data Analytics, Industrial Engineering, Robotics, or related field. Strong technical aptitude in data workflows, annotation, or process optimization. Analytical, detail-oriented, and eager to work in fast-paced, hands-on environments. Excellent communicator with a bias for ownership, quality, and continuous improvement.

The AI platform for human intelligence—we build models that rigorously vet PhDs and scale human data training for frontier LLMs. Our agent, Zara, autonomously sources and evaluates researchers and data trainers—powering post-training (RLHF, evals, red-teaming) for top AI labs. We've already built the infra layer that teaches language models how to think. Now, we're using that same layer to teach robots how to act.