Agent Orchestration
Multi-step agent workflows that plan, call tools, verify their own work, and recover — wired to run reliably in production, not just demo once.
Get started
Let's build something real.
I build production agent systems — agent orchestration, custom MCP servers, and the self-hosted runtimes that tie them together — for teams shipping with AI. The agent layer that takes you from demo to production, built by the person writing the code.
Use cases
What I work on
MCP Servers & Integrations
Custom MCP servers that connect agents to your APIs, data, and internal tools — including ones wrapping the Cursor, E2B, ElevenLabs, Twilio, and Smartlead APIs.
Autonomous Runtimes
Self-hosted harnesses that run your agents as CLIs, daemons, or bots — with config, audit trails, and recovery paths built in. The same shape as Agent AFK.
Prototype → Production
Taking an agent that works in a notebook and making it survive real inputs, real volume, and real edge cases once it goes live.
What people say
“What Griffin created is truly next-level. It makes generating leads and scheduling emails incredibly easy.”
“Griffin and the GRAIsol Team reduced our CPC by about 60% and marked up our click through rate up to 7.5%. The GRAIsol team is on some seriously next level use of AI and LLM modeling.”
“Agent AFK is my go-to now.”
Why me
Shipped MCP servers wrapping the Cursor, E2B, ElevenLabs, Twilio, and Smartlead APIs
Author of agent-framework — open-source orchestration for Claude Code
Built Agent AFK, a self-hosted runtime for autonomous agents
Have something gnarly to ship?
I take on a small number of agent-build engagements at a time. If you're past “we should use agents” and need someone who's actually shipped them to production, let's talk.