AI记忆系统研发专家/工程师（上海/北京/深圳）
Core
Design and develop an AI agent memory system that handles multiple memory types (core, episodic, semantic, procedural) and supports multi-modal inputs for downstream tasks.
Role type
Senior IC machine-learning engineer (agent memory systems)
Builds
Production memory systems for multi-agent collaboration
Domain
Artificial Intelligence / Large Language Models / Cognitive Science
Deliverable
production ML models
Required skills
Python, Go, C++, Agent memory architecture design, LLM agent frameworks (LangChain, LlamaIndex, Letta/MemGPT, AutoGen), Vector databases (Milvus, Qdrant, Faiss), Multi-modal processing, System design
Preferred skills
Cognitive science memory models, Conflict resolution algorithms, Dynamic weight adjustment strategies
Technologies
Python, Go, C++, LangChain, LlamaIndex, Letta, MemGPT, AutoGen, Milvus, Qdrant, Faiss
Responsibilities
Design agent memory systems covering various memory types; Build write, integrate, and retrieve pipelines for multi-modal inputs; Implement memory lifecycle management (decay, expiration, weighting); Develop hybrid retrieval systems combining vector and text search; Design multi-agent collaboration architectures; Support multi-modal memory input processing; Construct evaluation metrics for memory accuracy and effectiveness.