The newest and fastest-growing part of my work is building Generative AI tooling and autonomous agents that automate real work, not just demos.
What I’ve built
- Architected and deployed Generative AI tooling and autonomous agents using CrewAI to automate internal workflows and cut manual effort on the Customer Success team.
- Orchestrated LLMs into multi-step, tool-using workflows rather than single-prompt calls I make tools that get the work done, and help the team.
- Working with MCP servers and AI agent skills to give agents reliable access to tools and data, without handing them the keys to everything.
How I approach it
I treat agentic systems like any other production system: grounded in real data, guarded against the ways they fail (hallucination, prompt injection, runaway cost), and measured by whether they remove manual effort. Model choice is a deployment detail, so I keep it model-agnostic and route to whatever provider fits.
This site itself is heading toward a live demonstration of that. Coming soon is an AI chat you can interrogate. Until then, get in touch if you’d like to talk through a specific agentic build.