Engagements & pricing

Priced to the problem, not the clock

Three engagement shapes — advisory, build, retainer. No SaaS tiers, no per-seat games. We scope the problem first; the price follows the scope.

Every engagement starts the same way A free first conversation A written scope A number you agree to before work starts

Advisory

Hourly or short-term. For when you need a senior second opinion before you spend real money.

Book a consult

Hourly/ short-term

AI-readiness & agent-safety audits
Eval reviews — is your AI actually measured?
Architecture reviews & second opinions
Technical due diligence
Written findings you keep

Build Most common

Fixed-scope projects with fixed deliverables. The shape most work takes.

Scope a project

Fixed scope/ let’s scope it

price agreed in writing before work starts

Agent pilots: one workflow agentified end to end, with evals + human gates
MCP servers for your internal systems
RAG deployments — local-first available
Cloud migrations & architecture
Written architecture doc included
Monitoring & evals included
Documented handoff

Retainer

Monthly. For live systems and agent fleets that need an operator — watch, grade, improve.

Talk retainers

Monthly/ ongoing

Operate the fleet: agents, pipelines, dashboards
Monitoring & alerting, watched by a human
Evals drift review — quality graded month over month
Priority response
Monthly written report

Not sure which fits? Start with the free conversation — scoping is part of the job.

Compare engagements ↓

Your accounts, your code

Everything I build, you own outright

Work happens in your cloud accounts and your repositories. Local-first LLM work means your data can stay on your hardware. When the engagement ends, nothing is held hostage.

Book a consult How I work

Built in your AWS account, your repos, your infrastructure
Local-first models — private data never has to leave your hardware
Prompts, evals, and agent configs are deliverables too
Written architecture docs and runbooks stay with you
Monitoring dashboards your team can read without me
No lock-in — handoff is a deliverable, not an upsell

What each engagement can look like

Advisory: AI-readiness & agent-safety audit

You bring:

A stack that grew organically
A board asking "what about AI agents?"
A chatbot pilot nobody quite trusts

You get:

A written audit of what’s feasible now
An eval review of any AI already in production
A prioritized roadmap with cost ranges — and an honest "don’t build this" list

Build: agent pilot

You bring:

One workflow that eats staff hours
Systems with APIs (or at least screens)
A named human who can approve actions

You get:

That workflow agentified end to end
An MCP server exposing your system as tools
Evals, human gates, monitoring, docs, handoff

Retainer: operate the fleet

You bring:

Live agents or pipelines doing real work
No one who wants to be on call
A backlog of "someday" improvements

You get:

Monitoring & evals drift watched by a human
Steady iteration on the backlog
A monthly written report

Examples are illustrative shapes, not quotes. Every real engagement gets a written scope and a number before work starts.

Scope first, price second

The free first conversation and a short scoping pass produce a written scope with a number attached. You approve it before any billable work starts — no meter running during discovery.

Longer commitments, better terms

Retainers and multi-phase builds get priority scheduling and better rates than one-off sprints. If the work is ongoing, the pricing should be too.

Get in touch →

Compare engagements

What’s included, side by side

Included	Advisory	Build	Retainer
Getting started
Free first conversation	✓	✓	✓
Written scope before work starts	✓	✓	✓
Fixed price agreed up front	Hourly / day rate	✓	Monthly
Deliverables
Written architecture / findings doc	✓	✓	✓
Hands-on build	—	✓	Iteration on live systems
Monitoring & automated alerts	Recommendations	✓	✓ + watched
Runbooks & handoff docs	—	✓	Kept current
Ongoing
Response time	By appointment	During the project	Priority — usually within a business day
Continuous improvement backlog	—	—	✓
Monthly written review	—	—	✓
Work in your accounts & repos	✓	✓	✓

Qualitative by design — response times and inclusions are working practice, not contractual SLAs unless we put them in your scope.

Have a system that should exist but doesn’t?

Tell me what you’re trying to do. First conversation is free.

Book a consult See the work