I build the delivery system around your AI tools.
I work with a small number of engineering teams each year. These teams already use AI tools. They need better context, evaluations, review, and evidence.
What I help with
AI platform strategy
I turn tool licenses, experiments, and prompts into one delivery system that teams can use.
Context and MCP design
I give agents controlled access to the code, product, customer, and internal service information that they need.
Evaluation and evidence
I build tests and review records for your codebase. Leaders can see whether the work meets its goal.
Delivery workflow design
I design the steps for requirements, human decisions, automatic checks, and release approval.
How I work
Measurable results
The team must change how it works. My work reduces waiting, clarifies ownership, speeds up reviews, and creates an audit record.
Works with your tools
Your team can use Codex, Claude Code, Cursor, Copilot, or Windsurf. Tools change over time. The context, workflow, and verification methods remain useful.
Evaluations start early
I define success criteria and required evidence before the team starts implementation.
Your work controls the design
I study your team's work. Then I select agent methods, tools, and controls that support it.
What is different
I do the engineering work
I read the code, write the configuration, build the evaluations, and help release the first software change.
I test methods in my product
I build Squad at Remedys.ai. This work tests the methods that I use with clients.
Your team owns the result
Your team owns the system after the first release. I provide clear terms, instructions, and review methods that your leaders can continue to use.
Start with the workflow.
Email me about your product, the problem with AI adoption, your previous work, and the result that you need.