Lead Perception Engineer

Ovii's Interpretation of the Role

As Lead Perception Engineer at Origin, you will design and ship real‑time 3D perception pipelines that let autonomous construction robots understand complex sites. You’ll fuse LiDAR, stereo, IMU and camera data, build deep‑learning models, and ensure robust localization and calibration for edge deployment.

Role Snapshot

• Lead design of 3D perception pipelines
• Build and optimize deep‑learning models for scene understanding
• Develop sensor‑fusion and calibration systems
• Collaborate with navigation, manipulation and cloud teams
• Define and track perception KPIs
• Benchmark models for edge‑device performance

Must-Have Requirements

• C++ programming
• Computer vision fundamentals
• 3D perception algorithms
• LiDAR, Stereo, RGB camera processing
• IMU sensor processing
• Sensor fusion
• Localization / SLAM basics
• Deep learning (PyTorch, TensorRT)
• Calibration procedures (intrinsic/extrinsic)
• computer vision
• 3D perception
• sensor fusion
• deep learning

Nice-to-Have Signals

• Python programming
• Nvidia DeepStream / GStreamer / Holoscan
• ROS / ROS2
• OpenCV / PCL or similar libraries
• robotics projects
• ROS/ROS2
• Nvidia DeepStream

Work Setup

• Location: Bengaluru, India
• Work mode: ONSITE
• Employment type: Full-Time

Not Specified in JD

• Visa sponsorship
• Salary range
• Remote eligibility
• Education requirement
• Certifications
• Relocation
• Notice period
• Travel
• Security clearance
• Coding test
• Portfolio
• GitHub
• Writing sample
• Cover letter

What You'll Likely Work On

• Design, implement, and deploy real‑time 3D perception pipelines using LiDAR, stereo, RGB cameras and IMU.
• Develop algorithms for semantic scene understanding, ego‑motion estimation, and localization to feed downstream planning.
• Train, integrate and benchmark deep‑learning models for world understanding and surface‑finish classification, automating data collection and training pipelines.
• Create and maintain sensor‑fusion frameworks that combine classical methods with deep learning, handling noisy asynchronous data.
• Build robust calibration procedures (intrinsic/extrinsic) with drift detection and self‑healing capabilities, and define quantitative metrics for quality.
• Partner with navigation, manipulation and cloud teams to ensure perception outputs meet planning and fleet‑monitoring needs, tracking system‑level KPIs.

Good Fit If You Have

• Experience on robotics or vision‑based projects
• Familiarity with ROS/ROS2 or Nvidia DeepStream pipelines
• Ability to debug complex perception systems and manage large datasets
• Exposure to OpenCV, PCL, or similar computer‑vision libraries
• Comfort with edge deployment constraints