Robotics Perception Engineer

Auxo Talent; San Francisco (CA); On-site; USD 140,000 - 190,000; Full time

Job summary

Auxo Talent is seeking a robotics perception engineer to own real-time perception on real-world robots, from sensor integration to safety‑certified components.

You will build and optimize perception pipelines (cameras, LiDAR, IMUs), ensure low latency data flow to foundation models and control stacks, and collaborate with AI/ML teams.

Candidates should have 4+ years in robotics perception, strong C++ and Python, and hands-on sensor calibration; visa restrictions apply.

Qualifications

• 4+ years in robotics perception, computer vision, or a closely related field.
• Strong software engineering fundamentals in C++ and Python.
• Hands‑on experience integrating and calibrating sensors (cameras, LiDAR, IMUs, depth sensors).
• Experience building real-time perception pipelines on embedded or edge compute.
• Familiarity with streaming and networking systems for low-latency sensor data transport.
• Experience with ROS/ROS2 or similar robotics middleware, in production or research.
• Comfortable debugging across the full stack, from driver-level sensor issues to perception behaviour on live hardware.

Responsibilities

• Develop and maintain real-time perception pipelines, integrating sensors (cameras, LiDAR, IMUs) on physical robot platforms.
• Build and optimise object detection, tracking and scene understanding for safe operation in unstructured environments.
• Own the streaming infrastructure delivering sensor data to foundation models, teleoperation systems and onboard control stacks.
• Partner with the AI/ML team to connect perception outputs with learned policy inputs, ensuring low latency and high reliability.
• Implement and validate safety-certified perception components, including object detection and collision avoidance.
• Support bring-up and field testing of new camera hardware and sensor configurations.
• Contribute to system reliability, fault detection and recovery logic for robust perception in the field.

Skills

Robotics perception

C++ programming

Python programming

Sensor integration

Real-time systems

Tools

ROS/ROS2

TensorRT

Job description

A well-capitalised Physical AI company building a platform for complex, real-world environments. Think of them as the brain behind robots, from robot systems and infrastructure to the foundation models that control the robots. They're scaling fast across research and engineering.

Due to restrictions on visa transfers, we are unable to consider candidates who are not currently Green Card holders or US Citizens.

The Role

You’ll build and own the perception systems that give real-world robots real-time understanding of their surroundings. This means everything from sensor integration and calibration through to safety-certified object detection and streaming perception data to the foundation models and teleoperation stack. If you want your code to directly decide what a robot can see, understand and safely act on, this is that role.

What You’ll Do

• Develop and maintain real-time perception pipelines, integrating and calibrating sensors (cameras, LiDAR, IMUs, depth sensors) on physical robot platforms


• Build and optimise object detection, tracking and scene understanding for safe operation in unstructured environments


• Own the streaming infrastructure delivering sensor data to foundation models, teleoperation systems and onboard control stacks


• Partner closely with the AI/ML team to connect perception outputs with learned policy inputs, keeping data delivery low-latency and high-reliability


• Implement and validate safety-certified perception components, including object detection and collision avoidance, for real-world deployment


• Support bring-up and field testing of new camera hardware and sensor configurations


• Contribute to system reliability, fault detection and recovery logic for robust perception in the field

What We’re Looking For

• 4+ years in robotics perception, computer vision, or a closely related field


• Strong software engineering fundamentals in C++ and Python


• Hands‑on experience integrating and calibrating sensors (cameras, LiDAR, IMUs, depth sensors)


• Experience building real-time perception pipelines on embedded or edge compute


• Familiarity with streaming and networking systems for low-latency sensor data transport


• Experience with ROS/ROS2 or similar robotics middleware, in production or research


• Comfortable debugging across the full stack, from driver-level sensor issues to perception behaviour on live hardware

Nice to Have

• Background in multi-sensor fusion and calibration for real-world conditions


• Familiarity with teleoperation systems and their latency/reliability constraints


• Experience deploying and optimising perception models on edge or onboard compute (quantisation, TensorRT, etc.)


• Experience with humanoid or other legged robot platforms


• Prior work on early-stage hardware programmes (prototype or pre-production robots)