Founding ML Engineer - Robotic Policy Evaluation

One Robot (YC W26); San Francisco (CA); On-site; USD 150,000 - 275,000; Full time

Job summary

One Robot in San Francisco is hiring a researcher/engineer to advance its evaluation platform for robot manipulation policies. You will focus on training evaluation models, validating policy behavior on real deployments, and building scalable evaluation layers.

You will work with real customer data and drive improvements from data collection to production feedback, shaping how the eval system predicts and explains policy failures across long-horizon tasks.

Qualifications

• Strong Python and PyTorch coding skills.
• Experience training VLMs or LLMs.
• Experience developing and shipping evals for VLMs/LLMs.

Responsibilities

• Train evaluation models to classify and verify policy behavior.
• Build confidence layers translating outputs into actionable signals.
• Improve grounding of models for physical and spatial scenes.
• Develop data engine to sharpen eval models with deployments.

Skills

Python

PyTorch

VLM training

LLM training

Eval development

Job description

One Robot in San Francisco is hiring a researcher/engineer to advance its evaluation platform for robot manipulation policies. You will focus on training evaluation models, validating policy behavior on real deployments, and building scalable evaluation layers.

You will work with real customer data and drive improvements from data collection to production feedback, shaping how the eval system predicts and explains policy failures across long-horizon tasks.