— Consulting
Three ways to
work together.
Each engagement is fixed in scope and fixed in price. We talk first, I scope the work around your situation, and you see the exact price in the scoping document before any commitment. No hourly billing, no surprises. All three can evolve into longer partnerships once there is mutual trust and a proven track record.
AI Automation Sprint
From manual process to working AI system.
Most organisations have at least one process that consumes significant time, is largely repetitive, and could be automated with AI. The problem is that "we could automate this with AI" and "we have a working AI system in production" are separated by months of engineering that the demo never shows.
The AI Automation Sprint is a fixed-scope engagement. Week one identifies the highest-value process to automate. Then I design the appropriate AI architecture (not every problem needs 23 agents and a vector database), build and deploy it, and close with two weeks of production support to handle what surfaces when real users arrive.
What’s included
- Discovery and scoping session that identifies the right problem to solve, not just the stated one
- Architecture design document
- Full development and deployment
- Integration with existing tools and workflows
- Documentation and knowledge transfer
- Two-week post-delivery support window
Who this is for
Teams with a clear manual bottleneck, an AI feature that never made it to production, or a workflow that a demo proved could work but nobody built properly.
Architecture Review
Find what breaks before it breaks.
Most AI systems that fail at scale had design problems that were visible at the architecture stage and invisible to the people building them, who were too close to see them. The Architecture Review is a one-week engagement where I analyse your system from the outside: the data flow, the orchestration logic, the failure modes, the latency assumptions, and the scaling constraints.
The output is a written report of 20 to 40 pages with specific, actionable findings rather than general recommendations, followed by a two-hour session to walk through the findings and answer questions. Two weeks of async follow-up are included for clarifications.
What’s included
- Full technical analysis of your AI architecture
- Written report with prioritised findings
- Two-hour presentation and Q&A session
- Two weeks of async follow-up
- Optional: implementation roadmap for addressing identified issues
Who this is for
Teams about to scale an AI system who want a second set of expert eyes before committing. Startups preparing for technical due diligence. Organisations whose AI system has underperformed in production and who need to understand why.
Fractional AI Lead
Senior AI technical leadership. No full-time hire required.
A startup building a product with AI at its core needs senior architectural direction: someone who has seen these systems fail in production and knows how to prevent it. A full-time senior AI engineer or CTO costs €150,000 to €250,000 per year plus equity, and the hiring process alone takes three to six months.
The Fractional AI Lead retainer gives you the decisions that actually require senior judgement, without the salary, the equity, or the six-month search. We agree the monthly allocation up front, and I keep the number of concurrent retainers small enough that yours gets senior attention, not leftover attention. Strategic architectural decisions, not a warm body in a seat.
What’s included
- 20 to 40 hours per month of dedicated availability
- Weekly one-hour alignment session
- Architecture decisions and technical direction
- Code and design review for your engineering team
- Async availability throughout the month for time-sensitive questions
- Ongoing monitoring of AI performance and production health
Who this is for
Startups with an engineering team that needs senior AI direction. Scale-ups integrating AI into an existing product. Companies preparing an AI system for a first production deployment.
How this usually goes.
01
Conversation
A 30-minute call about what you are building and what is not working. No deck, no pitch. If I am not the right person, I will say so and suggest who might be.
02
Scoping
I write a short scope document with the problem as I understand it, the proposed approach, the deliverables, and the price. You either recognise your problem in it or you do not.
03
Work
Fixed scope, fixed price, agreed timeline. Weekly written updates. If the scope needs to change, we discuss it before it changes, not after.
04
Handover
Documentation, knowledge transfer, and a support window. The goal is that your team can operate what I built without me.
What I don’t take on.
Staff augmentation
I am not looking to be an extra pair of hands on a backlog. If the work does not require architectural judgement, someone cheaper should do it and you should let them.
Projects with no technical owner on the client side
If nobody on your team can maintain what I build after handover, the engagement produces a dependency rather than a system. I would rather train someone during the build than leave you stranded after it.
AI where AI is not the answer
A meaningful share of problems described to me as AI problems are data problems, process problems, or a missing integration. I will tell you that on the first call, and it costs you nothing.
Questions that come up.
Do you work with companies outside Spain and Portugal?
Yes. I work remotely across the EU and with US-based teams, and I am comfortable with the timezone overlap that requires. I am fluent in English, Spanish, and Portuguese, which removes most of the friction from European engagements.
What if my project doesn't fit neatly into one of these three?
Then we shape something that does fit. The three formats exist because they cover most situations, not because they are the only ones. Tell me the problem first and we work backwards from it.
Do you work on early-stage ideas or only production systems?
Both, but the engagement looks different. Early-stage work is usually an Architecture Review before anything is built, which is the cheapest moment to correct a design. Production work is usually a Sprint or a retainer.
How do I know you're the right person for this?
Read the Work page. If the problems described there resemble the problem you have, we are probably a fit. If they do not, we probably are not, and a 30-minute call will establish that faster than any proposal.