LLM Agent基础架构工程师-TRAE
Core
Design and build the core backend systems for TRAE, an AI coding agent that performs end-to-end software development tasks including code completion, generation, and execution.
Role type
Senior IC LLM Agent Infrastructure Engineer
Builds
AI IDE/Coding Agent backend, model access & scheduling infrastructure, Agent Runtime, and engineering platforms for context/prompt management and experimentation.
Domain
AI Software Development / LLM Infrastructure
Deliverable
production ML models
Required skills
Golang/Java/Python/C++, high-concurrency system design, LLM API engineering, model routing, streaming, function/tool calling, long-context handling, rate limiting, cost optimization, Agent loop optimization, task decomposition, context construction, failure recovery, tracing, A/B testing.
Preferred skills
AI IDE/Coding Agent experience, LLM Gateway, Agent Runtime, Prompt management platforms, LLM Tracing, commercial billing systems.
Responsibilities
Develop core backend systems for code completion, understanding, generation, and test execution; Build infrastructure for multi-model access, routing, and streaming; Construct engineering infrastructure for Agent runtime and context management; Optimize system performance for low-latency high-frequency requests and high-concurrency long tasks.
Seniority
Senior, hands-on IC
