AI/ML-Senior Perception Engineer

Job Requirements

Key Responsibilities:

Implement, validate, integrate and deploy robust perception solutions with sensor fusion for autonomous vehicles using LiDAR, Camera, Radar and GNSS sensors.

Elicitation of system requirements to derive ODD and KPI goals and instructions to capture datasets for training, validation and testing.

Develop and maintain intrinsic and extrinsic sensor calibration workflows

Build pre-processing and post-processing filters to clean and refine raw sensor data for downstream perception modules.

Optimize perception algorithms for real-world deployment, balancing accuracy, latency, and computational performance on embedded systems.

Test, validate, and fine-tune perception modules independently to handle edge cases, unusual driving conditions, and rare corner scenarios.

Independently design experiments, debug challenging issues, and verify solutions in simulation, test bench and on-vehicle tests.

Develop robust, production-ready C++ and Python code and integrate it into production pipelines

Collaborate with cross-functional teams to ensure smooth integration of perception, calibration, and localization functions with planning and control sub-systems.

Work Experience

Proven experience developing perception algorithms for autonomous vehicles or robotics covering object detection, semantic segmentation, tracking and localization using classic methods (e. g. RANSAC, Euclidean clustering, Occupancy Grid Mapping) and deep learning models (e. g. PointNet++, PointPillars) for Lidar point clouds and radar scans.

Hands on knowledge in sensor fusion techniques (Kalman Filters, Extended Kalman Filters, Unscented Kalman Filters, Particle Filters) for combining data from LiDAR, camera, radar, and GNSS.

Strong hands-on programming skills in C++ and Python.

Solid understanding of LiDAR, camera, RADAR, and GNSS sensor data, their characteristics and calibration mechanisms (intrinsic and extrinsic)

Knowledgeable in ROS2 software stack and Docker Compose tool for integration and deployment of perception algorithms.

Experience in implementing computational principles effectively using libraries like Eigen, Boost, GeographicLib etc.

Familiarity with object detection models (e. g. , YOLO, Faster R-CNN, SSD) and point cloud libraries (PCL).

Good understanding of real-time system constraints and optimization trade-offs for product-level deployment.

Ability to think independently, handle unusual edge conditions, and fine-tune algorithms for real-world robustness.

Excellent debugging and problem-solving skills.

Nice to Have:

Familiarity with SLAM, visual-inertial localization, or GNSS-INS integration.

Knowledge of parallel computing (CUDA, OpenCL, OpenMP) for real-time acceleration.

Prior experience with automotive-grade software development practices.