@jackjin1997
FreeBackend engineer focused on Agent Harness, Tool Use, MCP, Evals, and AI infrastructure.
FreeFree tier
About @jackjin1997
Zexu Jin (jackjin1997) is a backend engineer based in Hangzhou, China, with a strong focus on AI infrastructure, including Agent Harness, Tool Use, MCP (Model Context Protocol), Evals, and AI infrastructure. He maintains several prominent open-source projects such as langgraphjs (a framework for building language agent graphs), openclaw (a personal AI assistant), and hermes-agent (an agent that grows with the user), among others. His GitHub profile showcases active contributions to the AI agent and developer tooling ecosystem.
Key Features
Framework to build resilient language agents as graphs (langgraphjs)
Personal AI assistant for any OS and platform (openclaw)
Agent that grows with the user (hermes-agent)
Ultimate resource for OpenClaw: Skills, MCP Servers, and Agentic Intelligence (ClawForge)
Variance-aware benchmark for AI coding agents (AgentBench-Live)
Self-evolving Context Database for AI Agents (OpenViking)
Pros & Cons
Pros
- Active open source contributor with multiple starred projects
- Focus on cutting-edge AI infrastructure and agent development
- Pinned repositories cover diverse agent-related tools and frameworks
- Clear specialization in agent harness, tool use, and evaluation
Cons
- Limited information about production usage or enterprise support
- Projects are primarily in early stages (e.g., low star counts on some)
- No detailed documentation or tutorials provided on the profile
Best For
Building and deploying language-based AI agentsDeveloping personal AI assistantsEvaluating AI coding agent performance with variance-aware benchmarksCreating MCP servers and skill sets for agentic systemsManaging agent memory, knowledge RAG, and skills