Robot Perception Engineer

Location: San Mateo, United States; hybrid

Company

hirify.global builds data collection systems, large-scale data infrastructure, and software tools for training general-purpose robots.

What you will do

• Design human pose estimation pipelines for hand and full-body motion capture, teleoperation, and robot training annotations.
• Build camera calibration, hand-eye calibration, visual SLAM, and temporal alignment systems for multimodal robotic sensors.
• Train, evaluate, and optimize object detection, instance segmentation, and 6DoF pose estimation models.
• Deploy real-time perception inference on embedded robot platforms using TensorRT, CUDA, and custom kernels.
• Develop automated annotation and Auto QA pipelines to identify anomalous frames, failed demonstrations, and sensor dropouts.
• Collaborate with ML engineering and data infrastructure teams to connect perception outputs with VLA model training.

Requirements

• 5+ years of industry experience in robot perception or computer vision.
• Strong 3D vision fundamentals, including stereo cameras, structured-light cameras, and 3D reconstruction.
• Experience with SLAM or V-SLAM development and human pose estimation for hand or full-body joints.
• Experience training, tuning, and evaluating deep learning perception models.
• TensorRT deployment and real-time inference optimization experience on embedded platforms such as Jetson or Horizon.
• Proficiency in C++, Python, CUDA, ROS or ROS2, and AI coding agents.

Nice to have

• Experience with 6DoF object pose estimation, 3D Gaussian Splatting, or NeRF.
• Experience with robot manipulation, teleoperation, automated annotation, or ground-truth generation systems.
• Published research in perception, pose estimation, or robotics.

Culture & Benefits

• Work on robot training data and embodied intelligence.
• Collaborate with robotics engineers and ML researchers.
• Use proprietary humanoid robots, camera arrays, and data gloves.
• Operate in a fast-paced, high-autonomy 0→1 environment.