Service
Agentic Development
Building software with AI coding agents as real members of the team, from the first specification to a merged, reviewed and deployed change.
Agentic development is more than giving developers a chat window. It is a different way of working, in which agents take on real tasks: exploring a codebase, writing and refactoring code, running tests and preparing changes for review. Done well, it frees people to focus on intent, design and judgement.
I help teams make that shift in their own codebase, on their own work. Not as a demo, but as a way of working that lasts: with clear roles for people and agents, a repository that agents can understand, and a delivery flow that stays safe as the pace goes up.
How I help
Start with real work
We pick real tickets from your backlog and solve them together with agents, so the team learns on work that matters.
Shape the codebase
Agents copy what they find. I help make the repository agent-ready, with context, conventions and tests that guide them well.
Keep people in charge
Agents propose, people decide. We agree on where autonomy helps and where a human always has the final say.
What working together looks like
- A short exploration of your codebase, team and current way of working
- Hands-on pairing sessions with developers and agents on real tasks
- Context and rules files that teach agents how your team works
- A delivery flow of review, testing and deployment that keeps up with the new pace
- Gradually handing over, until the team works this way on its own
Questions to start with
- Which parts of your daily development work feel most repetitive?
- How much does your team already use AI tools, and how?
- What would you want to stay firmly in human hands?