大模型Agent平台研发工程师-Data AML
Core
Build a core inference outsourcing platform and Agent infrastructure to support automated model access, high-performance inference, and intelligent agent task management.
Role type
Senior IC machine-learning engineer (inference & agent platform)
Builds
Inference outsourcing platform, Agent state management systems, developer toolchains, and hybrid runtime environments
Domain
Artificial Intelligence / Large Language Models / Distributed Systems
Deliverable
production ML models
Required skills
Python, C/C++, Linux, Distributed Systems, Large Model Inference, Transformer Architecture, vLLM, TensorRT-LLM, High Concurrency Systems, Redis, PostgreSQL, Vector Database, Agent Frameworks (LangGraph, AutoGen, Dify)
Preferred skills
Speculative Decoding, PagedAttention, Hardware-Software Co-design, OpenAI Assistants API, MCP Protocol, SDK Development, RLHF Optimization Algorithms
Technologies
Hugging Face, vLLM, TensorRT-LLM, Redis, PostgreSQL, Vector Database, LangGraph, AutoGen, Dify, MCP
Responsibilities
Design and implement cloud storage and state machine systems for agent task rollback and debugging; Develop observability systems for full-link tracing of agent decision trees; Create SDKs for seamless integration of local and cloud tools; Deploy advanced inference capabilities like automatic speculative decoding and quantization.
Seniority
Senior, hands-on IC
