Services

Security baseline for small teams

Who it's for: teams without a security person, who know they should look at security but don't know where to start. Often it comes up because a customer has sent a security questionnaire, or because the team is about to raise money.

What I do: I check the places where small teams most often get hurt. That covers who has access to what, how you handle secrets, your CI and deploy identities, your dependencies, your logging, and what you would do if someone reported a problem tomorrow.

What you get: a short list of findings in priority order, written so your team can act on it. I can stay on to help fix them. This isn't a penetration test - it's the groundwork that makes one worth paying for later.

Agentic engineering setup

Who it's for: teams already using coding agents like Claude Code or Codex, who find the results uneven. Some work is great, some is a mess, and nobody is sure what the agent is allowed to touch.

What I do: I set up the workflow in your repos: how work is planned, what the agent can and can't change, and the checks that run before anything merges. Rules are enforced by tooling and CI, not just written down in a prompt.

What you get: repos where agents do useful work inside clear limits, the skills and guardrails written down for your team, and a team that knows how to change them.

Agentic codebase recovery

Who it's for: teams that relied too much on unrestrained agents to get a product out, and now can't work with the code. Bugs come back after they're fixed. Small changes break things somewhere else. Nobody is sure how it all fits together, and asking the agent to fix it makes it worse.

What I do: it starts with a fixed-scope assessment. I read the code - actually read it - and work out how it behaves, where it's fragile, and what it would take to make it solid. Then we pick a path:

• Untangle and fix. I find the root causes, add tests around what matters, and make the code stable enough that your team can change it again.
• Rebuild with a proper harness. When fixing would cost more than starting over, I rebuild the weak parts with agents working inside clear limits: planned work, rules enforced by tooling, and checks that run before anything merges.
• Ongoing human review. I review the pull requests your agents and team produce, as an engineer who still reads and reasons about code line by line, so problems get caught before they ship.

What you get: an honest written account of where your code stands and which path makes sense, then a product your team can trust and keep building on.

Hands-on build help

Who it's for: teams with a hard feature or a whole product they need built, and no one senior enough free to own it.

What I do: I work with you to define a fixed scope, then design, build, test, and ship it inside your codebase and your process. I've built and run software end-to-end for over 25 years, from nuclear inspection tools to cloud platforms.

What you get: working software in production, code your team can maintain, and a handover so it doesn't depend on me.