Ways I can help
Use AI coding tools without losing control
Practical help getting real engineering value from AI-assisted development, coding agents, and agentic workflows - without the discipline your codebase depends on quietly disappearing.
Let's look at how you're using AIYou might come to me when...
Our developers are all using AI differently, with no shared approach.
We're shipping faster, but I'm worried we're creating new problems just as fast.
We want to use coding agents seriously, without losing engineering discipline.
Nobody has really set up the guardrails, context, or review habits this needs.
I want an honest opinion on how well this is actually working for us.
Where I can help
- Coding-agent workflows, including orchestrating multiple agents or tools on one task
- Repository and project context for AI tools
- Review and testing practices
- Guardrails for what agents should and shouldn't touch
- Documentation as something AI can actually use
- Whether AI is actually improving delivery, not just activity
How this usually starts
Often this starts with a short look at how the team is actually using these tools today - what's working, what's quietly getting risky - and then some practical changes to workflow, review, and documentation rather than a big new process.