Sr. Perception Engineer

Embedding VC•San Francisco, CA

About The Position

We are seeking a Senior Perception Engineer to develop advanced perception algorithms and systems for next-generation warehouse automation robots. The role focuses on enabling robust 3D perception from cameras and LiDAR data for object understanding, scene reconstruction, and manipulation in unstructured environments. You will work across the perception stack from sensor calibration and data processing to algorithm design and real-time deployment.

Requirements

Bachelor or above in Robotics, Computer Vision, or related field.
5+ years (including Master / PhD) of professional experience in 3D perception for robotics or autonomous systems.
Over two years of experience leading and delivering high-impact perception initiatives.
Proficiency in point cloud processing and geometry-based vision (e.g., PCL, Open3D, ROS2 perception stack).
Strong software development skills in C++.
Demonstrated experience deploying perception systems in real robotic environments.
Startup mindset: hands-on, proactive, and ownership-driven.

Nice To Haves

Familiarity with AI-based perception or deep learning frameworks (PyTorch, TensorFlow).
Experience with multi-sensor fusion, calibration, and synchronization.
Knowledge of GPU optimization and real-time inference (TensorRT, ONNX Runtime).
Experience with different simulation environments (Isaac Sim, Gazebo).
Background in warehouse or industrial robotics.

Responsibilities

Develop algorithms for 3D perception tasks such as object detection, pose estimation, and scene reconstruction using multiple RGB-D and LiDAR data.
Build and maintain perception pipelines that combine geometric methods with data-driven approaches.
Ensure reliable and efficient real-time performance on Industrial PC platforms.
Contribute to processes for data collection, labeling, evaluation, and model training with consistent turnaround times.
Collaborate with motion planning, manipulation, and system software teams for seamless integration into autonomous workflows.