AI decisions, made by someone who makes them every day
Your team is already using AI. The question is whether anyone owns it: the roadmap, the tooling choices, the custom agents that keep working after the demo. I do this as a fractional architect for engineering orgs of 20–100 people and founder-led product companies. You get a staff-level engineer's judgment without the staff-level hire, and I write code, not slideware.
Email me and we'll talk. This isn't an engagement you book from a calendar link.
Four things, owned rather than advised on
The AI roadmap
A prioritized, sequenced plan for where AI belongs in your product and your process, revisited as the ground shifts. Not a strategy deck: a living document with owners and dates.
Custom agents and integrations
Built, shipped, and maintained. The maintenance clause matters: an agent nobody maintains is a demo, and demos rot. What I build, I keep working.
Vendor and tooling review
Which coding agents, which models, which harness, which of the seventeen vendors pitching you this quarter. I use these tools daily on my own projects, so my opinions come from commit history, not analyst reports.
Team enablement
Getting your engineers productive with AI-assisted development: workflows, skills, review practices, and the guardrails that keep quality up when velocity jumps.
The shape of the engagement
A recent engagement of this shape: fractional architect for a founder-led product company building an AI-assisted application for a specialized consulting market. I owned the architecture decisions, evaluated and selected the agent tooling, built working spikes to de-risk the hard parts, and kept the roadmap honest about what the technology could actually do this quarter versus next year.
That's the pattern: regular working sessions, real code in the repo, and one person accountable for the technical direction of the AI work. It runs alongside your team, not instead of it.
I've been building software professionally since 1998, most recently as an Engineering Manager and Architect at Codingscape, and before that running technology teams at RevUnit and Credit One Bank. I still ship my own software, including two iOS apps on the App Store, and I write about AI-assisted engineering at georgediab.com.
A fit when
- You have 20–100 people, or you're a founder building a product with a small technical team
- AI matters to your roadmap but doesn't justify a full-time architect hire yet
- You want working code and maintained systems, not a recommendations document
- Someone senior needs to own the tooling and vendor decisions
Not a fit when
- You need a large delivery team on day one. I start as one architect; when the work genuinely needs more hands, I bring in one or two AI-native engineers, scoped and priced before anyone starts
- You want a one-time training workshop; that's a different, smaller engagement, and I'll say so
- Nobody on your side can spend an hour a week with me; the retainer works through your team, and without that contact it stalls
Tell me what you're building and where it hurts
The first conversation is free and useful on its own: you'll leave with my honest read on your situation, including "you don't need me" if that's the truth. Retainers start at $3,000 a month and get scoped to your team before anything is signed.
From $3K/mo · scoped before signing · no long-term lock-in