[6 months, 4-5 days/week] Name of Project: Autonomous Driving Agent System Efficiency & Quality Enhancement Objectives: -The project aims to systematically identify and optimize bottlenecks in the Autonomous Driving Agent (decision-making and planning system) through data-driven methods, significantly improving its operational efficiency and stability. The intern's contribution will directly accelerate the iteration cycle of core algorithms and have a measurable impact on reducing system computational load and improving traffic flow efficiency. Main Tasks: -Utilize large-scale real-world and simulation data to analyze Agent performance bottlenecks. Develop data analysis scripts and visualization tools to pinpoint root causes. -Build and enhance Retrieval-Augmented Generation (RAG) systems for autonomous driving scenarios, improving the accuracy and efficiency of knowledge retrieval and decision support. -Participate in the post-training (SFT, RLHF) of large language models and develop LLM-based agent applications for task planning and decision-making in driving scenarios. -Conduct regression testing and performance evaluation of the optimized Agent systems, ensuring stability and efficiency improvements Key Qualifications: -Proficient in Python programming and major deep learning frameworks (e.g., PyTorch, Transformers). -Hands-on experience in RAG system development, including retrieval algorithms, vector databases, and prompt engineering. -Familiar with LLM agent application development frameworks (e.g., LangChain, LlamaIndex). -Experience with large-scale model training/inference optimization and related libraries (e.g., vLLM, DeepSpeed). -Knowledge of autonomous driving systems (Perception, Prediction, Decision-Making, Planning). -Publications or projects in NLP/LLM-related field Language: Chinese: Conversational English: Basic Knowledge Education: Degree: Bachelor or above Major: Computer Science