- 针对端到端和 VLA 自动驾驶模型，开展强化学习算法研究与创新。

- 探索先进的强化学习方法（如策略优化、在离线 RL、层次化 RL、多智能体 RL），提升算法的鲁棒性和泛化能力。

- 探索先进的强化学习方法在自动驾驶领域的应用和拓展。

- 设计评测基准，并在复杂动态驾驶场景中开展实验验证。

- 与感知、规划、仿真团队协作，将 RL 方法集成到端到端自动驾驶训练框架中。

- 形成研究成果，撰写技术文档或对外发表技术报告。

- Research and develop novel reinforcement learning algorithms for end-to-end autonomous driving and VLA (Vision-Language-Action) models.

- Explore advanced RL techniques (e.g., policy optimization, online/offline RL, hierarchical RL, multi-agent RL) to improve robustness and generalization.

- Familiar advanced RL algorithm(e.g. GRPO, GSPO etc.) to improve autonomous driving.

- Design benchmarks and conduct experiments to evaluate algorithm performance in dynamic driving environments.

- Collaborate with perception, planning, and simulation teams to integrate RL methods into E2E autonomous driving pipelines.

- Publish technical reports and document research findings.

1. 计算机、机器学习、自动化、机器人等相关专业硕士或博士学历。

2. 深入理解强化学习、深度学习与控制理论。

3. 熟悉 PyTorch/TensorFlow 及常用 RL 框架（如 RLlib、Stable-Baselines3）。

4. 在自动驾驶、机器人或多智能体系统方向有研究经验者优先。

5. 具备扎实的 Python/C++ 编程能力，熟悉大规模训练。

6. 具备较强的分析与问题解决能力，能够独立开展研究。

7. 具备良好的英文读写能力。

1. Ph.D. degree in Computer Science, Machine Learning, Robotics, or related fields.

2. Strong knowledge of reinforcement learning, deep learning, and control theory.

3. Hands-on experience with PyTorch/TensorFlow and common RL libraries (e.g., RLlib, Stable-Baselines3, etc.).

4. Research background in autonomous driving, robotics, or multi-agent systems is highly preferred.

5. Solid programming skills in Python/C++, good understanding of large-scale training.

6. Strong analytical and problem-solving skills; ability to work independently.

7. English reading/writing proficiency.