Algorithm Expert

End-to-End Model Design & Innovation: Lead the architecture design, implementation, and optimization of next-generation End-to-End autonomous driving models, covering the full stack from perception, prediction, decision-making, to planning.
Generative AI Exploration: Develop and implement E2E algorithm solutions based on cutting-edge generative technologies, including VLM/VLA (Vision-Language/Action Models), LLMs (Large Language Models), and Diffusion Models.
Inference Optimization & Deployment: Spearhead model inference acceleration and performance optimization. Drive model lightweighting strategies for embedded platforms to resolve high latency and compute bottlenecks in real-world applications.
Data Engine & Evaluation: Participate in the development of automated data pipelines and establish robust model evaluation metrics. Continuously iterate and enhance E2E model performance through closed-loop data mining.
Cross-functional Collaboration: Collaborate closely with system architects and embedded engineering teams to ensure seamless model integration and deployment. Solve practical performance issues and continuously improve model effectiveness in actual road tests.
Frontier Tech Research: Stay at the bleeding edge of large-scale models and reinforcement learning (RL). Explore the migration and application of these advanced technologies within commercial vehicle driving scenarios.
Education: Master’s degree or PhD in Computer Science, Artificial Intelligence, Information Engineering, Electronic Engineering, Robotics, or a related field.
Experience:
3+ years of hands-on experience in deep learning algorithm development, preferably within the autonomous driving or robotics industry.
A strong publication record in top-tier AI/CV conferences (CVPR, ICCV, NeurIPS, etc.) or proven success in relevant open-source projects is highly preferred.
Technical Expertise:
Proficient in Python and C++, with extensive practical experience in deep learning frameworks such as PyTorch or TensorFlow.
Deep understanding of mainstream algorithms in perception, prediction, and planning/control.
Solid grasp of Imitation Learning (IL) and Reinforcement Learning (RL) principles.
Strong experience in model inference optimization, including model compression techniques (pruning, distillation) and quantization (INT8/FP16). Familiarity with high-performance inference frameworks like TensorRT, ONNX Runtime, or vLLM is highly desirable.
Problem Solving: Excellent logical thinking, systematic problem-solving abilities, and strong fundamentals in data structures and algorithms. Passionate about autonomous driving technology and eager to tackle challenging real-world engineering problems.
Communication: Fluent in English and Mandarin, capable of effectively collaborating with global HQ and local cross-functional teams.

Employment type
Full-time

Work arrangement
No