# Fractional AI Architect

You are about to hire for a problem you cannot specify yet.

Ongoing architecture, evaluation, and stack decisions with your CTO — at eight, twelve, or twenty hours a week.

3 months minimum · $9K-15K/mo

Three tiers: 8, 12, or 20 hours per week. Three-month minimum, because anything shorter is a project, not a retainer.

## Who this is for

- You are shipping AI features and the architectural decisions are being made by whoever is closest to the ticket.
- You have an open req for a senior AI platform hire and no confident spec for it.
- You want someone accountable for the technical direction without adding permanent headcount.

## Who this is not for

- You need an implementer. This is architecture, review, and decisions — hands stay on the keyboard for prototypes and spikes, not for delivering your roadmap.
- You want a name for a deck or a diligence round. This is hands-on direction, and it earns its place in the architecture reviews, not on a slide.

## What you get

- Architecture ownership: the decisions get made, written down, and defended.
- Evaluation and SLO frameworks that give your team a shared definition of working.
- Stack and vendor selection with the trade-offs written out, not asserted.
- Code and design review on the parts that matter, at the depth the repos on this site should suggest.
- Hiring specification: what to hire for, in what order, and what to stop paying for.

## How it runs

### Month 1: Read and rank

The whole system read in one pass, decisions inventoried, and the top three architectural risks named and ranked.

### Month 2: Standards

Evaluation harness, SLOs, cost controls, and review standards in place, so quality stops being a matter of opinion.

### Month 3+: Cadence

Weekly working sessions, design review, and whatever the roadmap turns out to need. Renewed monthly, cancellable with notice.

## Reported experience

- 6+: Platforms taken from prototype to production
- 71% to 96%: SLO compliance improvement across three client platforms
- ~40%: Release-cycle overhead reduction across four distributed teams

## Evidence and limits

Anonymized client outcomes (author-reported, reviewed 2026-09-22): Reported by Nick Cerutti; client identities and raw operational records are not public. $200K/year is aggregate savings across engagements. $40K/day was projected exposure; 72 hours describes that incident, not a response or resolution guarantee. Zero-known-incident statements describe observations, not proof of absence.

[Source](https://nickcerutti.com/work)

## Questions

### Why a three-month minimum?

Because the first month is spent reading the system. A one-month retainer buys you the diagnosis and none of the value.

### Should we just hire someone full time?

Probably, eventually. The market median for an AI governance or platform hire runs around $169K plus equity and three to six months of search. Three months of fractional costs less than the search and tells you exactly what to write in the req.

### How much do you actually write code?

Enough to stay honest — prototypes, spikes, and the occasional hard piece. Not enough to become a dependency. If your roadmap needs an implementer, that is a project engagement.

### What happens if it is not working?

Cancel with 30 days' notice. There is no lock-in and no termination fee. A retainer that has to be enforced by contract is already over.

## Contact

[Discuss this engagement](mailto:n@nickcerutti.com) · [Pricing and terms](https://nickcerutti.com/pricing)

Canonical page: https://nickcerutti.com/services/fractional-ai-architect