豆包大模型Agent算法工程师 - 火山方舟
Core
Design and optimize the architecture and runtime mechanisms for managed AI agents and harnesses within a large-scale model platform.
Role type
Senior IC Agent Algorithm Engineer (Harness & Multi-Agent Systems)
Builds
Managed-Agent platform services, including harness runtime, tool integration, and multi-agent collaboration frameworks.
Domain
Large Language Models (LLM) and AI Agent Systems
Deliverable
production ML models
Required skills
Agent Harness architecture design, Multi-Agent collaboration scheduling, LLM API integration, ToolUse and FunctionCalling implementation, RAG retrieval techniques, Agent evaluation framework construction, Code evolution and AutoPE, ReAct reasoning patterns, KV Cache optimization.
Preferred skills
Experience with AlphaEvolve, Hyper-Agents, MCP (Model Context Protocol), A2A protocols, and deep practical usage of commercial Agent products.
Responsibilities
Select and optimize technical architecture for the harness system, enhance built-in tools and context mechanisms, build end-to-end evaluation systems for agent iterations, and define frontier innovations for model-harness co-evolution.
Seniority
Senior, hands-on IC