1. Design, develop, and optimize multi-agent architectures to enhance capabilities in planning, tool usage, instruction following, RAG, knowledge learning, demand understanding, environment interpretation, human-computer interaction, and multi-agent coordination.
2. Undertake machine learning, data analysis, and programming tasks; construct related benchmarks and enhance their long-term effectiveness from first principles.
3. Design, develop, and experiment with self-play mechanisms and reward models to enable agent self-evolution through closed-loop systems.
任职要求
1. Master’s degree or above in Computer Science, AI, Software Engineering, Mathematics, or related fields (Bachelor’s acceptable for outstanding candidates).
2. Strong Python programming ability with solid algorithmic skills and clean coding practices.
3. Familiarity with at least one deep learning framework (PyTorch, TensorFlow, etc.); open-source experience and large model fine-tuning are preferred.
4. Strong critical thinking, accountability, open-source mindset, good collaboration, and problem-solving skills.
5. Preferred:
- Publications in top conferences like ICLR/ICML/ACL/EMNLP/NeurIPS on LLM or agent-related topics.
- Contributions to open-source projects or independent AI agent development experience.
- Familiarity with Reinforcement Learning (RL) is a plus.
