Assistant Algorithm Engineer

Assistant Algorithm Engineer

Job Responsibilities:

1. Responsible for the design, implementation, and optimization of global/local path planning algorithms (such as A, Dijkstra, RRT/RRT, DWA, Lattice, Hybrid A*, APF, etc.) to meet real-time and robustness requirements in dynamic environments.

2. Solve path planning challenges in complex scenarios (e.g., obstacle avoidance in narrow passages, dynamic obstacle prediction, multi-robot collaborative planning, 3D space planning). Explore cutting-edge technologies, such as end-to-end planning algorithms based on deep learning (CNN/GNN) and reinforcement learning.

3. Optimize path smoothness, safety, and energy efficiency while supporting multi-objective constraints (e.g., shortest time, lowest energy consumption, minimal risk).

4. Collaborate with perception, control, and simulation teams to integrate and debug planning algorithms with sensor data (LiDAR, cameras, IMU).

5. Develop behavioral decision-making in low-speed scenarios, embed traffic rules (e.g., lane changes, intersection navigation), and design emergency obstacle avoidance strategies. Ensure safe planning in human-robot collaboration scenarios and multi-robot cooperative path allocation.

Qualifications:

1. Master’s degree or higher in Computer Science, Automation, Aerospace, Geomatics, or related fields. Bachelor’s degree applicants must have at least 3 years of relevant project experience.

2. Proficiency in C++/Python and familiarity with Linux development environments. Experience with open-source frameworks such as ROS/ROS2, PCL, and OpenCV.

3. In-depth understanding of algorithms like A, RRT series, DWA, etc., with experience in algorithm improvement (e.g., RRT-Connect, Informed RRT). Familiarity with trajectory optimization methods (polynomial interpolation, spline curves, optimal control) and feasibility handling under kinematic/dynamic constraints.

4. Hands-on experience in practical path planning projects (e.g., autonomous driving decision-making modules, UAV trajectory planning, warehouse robot scheduling systems).

5. Understanding of interface requirements for perception (object detection, tracking) and control (trajectory tracking) modules, with the ability to integrate and debug systems.

6. Knowledge of reinforcement learning (PPO, DQN) applications in planning or experience with CUDA-accelerated algorithm development is preferred.