Perception AI Researcher for Real-World Robotics

Precision Labs; San Carlos (CA); On-site; USD 200,000 - 300,000; Full time

Benefits offered by this job

Medical coverage

Dental coverage

Vision coverage

Generous PTO

401(k) match

FSA/HSA options

Commuter benefits

Disability insurance

Life insurance

EAP

Onsite snacks

Job summary

1X is building humanoid robots for home use, and we are seeking a perception engineer to own perception from prototype to fleet deployment. You will design and ship vision systems that measure geometry, recover pose, and reconstruct terrain and scenes that NEO must walk through and manipulate.

You will drive calibration, 3D scene understanding, and sensor fusion across cameras and inertial sensors, ensuring data is spatially and temporally consistent.

Qualifications

• Proven track record shipping perception systems on real hardware.
• Train, evaluate, and deploy deep learning models for detection, depth, or pose estimation in PyTorch.
• Geometric computer vision and camera calibration expertise.
• Own capability end-to-end from problem to deployment.
• Strong Python and modern C++ in large codebases with testing and code review.

Responsibilities

• Own perception for NEO from prototype to fleet deployment.
• Build vision systems to measure geometry, recover pose, and reconstruct terrain.
• Ship pose estimation models for large‑scale data labeling and motion generation.
• Drive sensor fusion across cameras and inertial sensors for consistent data.
• Calibrate cameras at scale and monitor drift across the fleet.
• Establish datasets and metrics to show perceptual improvements in robot behavior.

Skills

Python

C++

Deep learning

PyTorch

Geometric vision

Camera calibration

End-to-end ownership

Production ML

Testing & review

Robotics experience

Education

Advanced degree in CV/Robotics

Tools

PyTorch

Job description

1X is building humanoid robots for home use, and we are seeking a perception engineer to own perception from prototype to fleet deployment. You will design and ship vision systems that measure geometry, recover pose, and reconstruct terrain and scenes that NEO must walk through and manipulate.

You will drive calibration, 3D scene understanding, and sensor fusion across cameras and inertial sensors, ensuring data is spatially and temporally consistent.