Perception Engineer - Apprentice

About Origin Origin (previously 10xConstruction) is building general-purpose autonomous robots for US construction to tackle rising costs, safety risks, and labour shortages. Our modular, multi-trade platform combines purpose-built hardware with real-time site intelligence to navigate complex environments and execute tasks with precision. Trained in high-fidelity simulation and already deployed on live sites, our robots deliver 5x faster execution, 250%+ margin expansion, and significant cost savings. Join India’s most talent-dense robotics team consisting of individuals from IITs, Stanford, UCLA, etc. About the role As a Perception Engineering Intern / Apprentice at Origin (Formerly 10xConstruction), you will help our autonomous drywall-finishing robots "see" the job-site. You'll design and deploy perception pipelines—camera + LiDAR fusion, deep-learning vision models, and point-cloud geometry—to give the robot the awareness it needs. Key Responsibilities Build ROS 2 nodes for 3-D point-cloud ingestion, filtering, voxelisation and wall-plane extraction (PCL / Open3D). Train and integrate CNN / Transformer models for surface-defect detection and semantic segmentation. Implement RANSAC-based pose, plane and key-point estimation; refine with ICP or Kalman/EKF loops. Fuse LiDAR, depth camera, IMU and wheel odometry data for robust SLAM and obstacle avoidance. Optimize and benchmark models on Jetson-class edge devices with TensorRT / ONNX Runtime. Collect, label and augment real & synthetic datasets; automate experiment tracking (Weights & Biases, MLflow). Collaborate with manipulation, navigation and cloud teams to ship end-to-end, production-ready perception stacks . Solid grasp of linear algebra, probability and geometry; coursework or projects in CV or robotics perception. Proficient in Python 3.x and C++17/20; comfortable with git and CI workflows. Experience with ROS 2 (rclcpp / rclpy) and custom message / launch setups. Familiarity with deep-learning vision (PyTorch or TensorFlow)—classification, detection or segmentation. Hands-on work with point-cloud processing (PCL, Open3D); know when to apply voxel grids, KD-trees, RANSAC or ICP. Bonus: exposure to camera–LiDAR calibration, or real-time optimization libraries (Ceres, GTSAM). Preferred Skills: Python 3.x C++17/20 ROS 2 PyTorch Open3D RANSAC For apprentice/Intern : This is a six-month contract-to-hire position, with potential for longer-term opportunities depending on your contributions and interests. We’re excited to work with motivated individuals who are eager to support real-world robot development and grow alongside us.