AI consulting for teams that already have constraints
I am an independent AI solutions architect and enterprise AI consultant based in Madrid. I work with teams in Spain and across Europe (remote by default) on systems that have to fit existing processes, tools, and owners, and still run after the demo.
If you are looking for a freelance AI solutions architect or an AI consultant in Madrid who will tell you when a model is the wrong tool, this is the page.
What I work on
- Agent workflows with a clear trigger, owner, and fallback.
- Retrieval systems that return the source, page, or record behind an answer.
- Internal tools that remove a specific manual step from an existing process.
- Integrations with the systems where the team already works.
- Deployment, evaluation, logs, and failure handling.
- Technical workshops that force a decision: ship it, wait, or drop the idea.
When I am useful
A team has a concrete workflow but does not know whether it needs a model, a deterministic rule, or both.
A prototype works in a demo and now needs evaluation, permissions, logs, and an owner.
Useful knowledge is spread across documents and systems, but an unsupported answer would be worse than no answer.
The team needs a technical view before it commits to a vendor or a long build.
How I work
- Pick one workflow and name the person responsible for its outcome.
- Map the inputs, systems, decisions, and failure cost.
- Define examples that separate a useful result from a confident mistake.
- Build the smallest version that can answer the real question.
- Add the deployment path, logs, fallback, and handover needed to run it.
What I will not sell
A chatbot wrapper presented as a new operating model.
A prototype with no owner, no data path, and no fallback.
A model in the loop for a decision that already has a rule and a named person.
A vendor comparison that exists to delay a choice the team can already make.
If the honest answer is that the process should stay deterministic, I will say so.
Background (for hiring context)
I was a Solutions Architect at Multiverse Computing in Madrid from November 2025 to July 2026 (CompactifAI pre-sales and PoCs). Before that I was a Data Scientist at AILY LABS (July 2023 to November 2025), including production LangChain agents and Graph RAG over factory manuals. I am not currently employed by Multiverse.
Public notes on selected work: case studies. For a regulated-setting assessment note (not a claim of pharma client delivery): AI for pharmaceutical operations.
Contact
Email hello@felipebasurto.com with the workflow, the constraint, and what would count as a useful result. Madrid-based. Spain and Europe. Remote is the default.