Training · Mentoring · Projects
Training and guidance for leadership and technical teams: what AI can do today, where it fails and how to take it into production on your own terms.
Leaders need judgement to make decisions about AI. Builders need a method to take it into production. We work with both, and best of all in the same room.
The mental model to assess proposals, talk to your technical teams as an equal and tell what's real from what's hype. No coding.
Specs before code, agents with well-designed scaffolding and evaluations from day one: the craft of building with AI, beyond picking a model.
Where a bad call costs safety, money or reputation, AI comes in as a copilot, not an autopilot: knowing where it fails and where a person decides.
Training for teams, one-to-one support or a project guided end to end. Always on your real cases.
Six weeks with a track for leaders — the map of AI and the decisions it demands — or for technical teams: specs, agents, evaluations and the AI-assisted product cycle.
Three months on your real agenda, whether you lead or build. From the first session you leave with something that changes how you work, and by the end you carry on by yourself.
We guide your team through an AI software project end to end, so the know-how stays in-house. When it makes sense, we build it ourselves too.
Practice first, theory behind it: the concepts come in while we work on what's yours.
You tell us about your context, your team and where you feel friction. We listen. No cost and no commitment.
We come back with a route built for you: which topics to explore and in what order, based on what you told us.
Every session happens in your own tools, on real cases, and ends with one concrete action for that same week.
At the close we review what changed and what comes next, and stay available for the questions that come up later.
Andrés Pérez is an electrical engineer from Pontificia Universidad Católica de Chile and holds an M.Sc. in Telecommunications Engineering from Politecnico di Torino. Since 2016 he has taken artificial intelligence systems into production in biotech, aviation, retail and finance, and now puts that craft into helping those who lead and those who build use it with judgement.
Multi-agent assistants and language models taken into production in large-scale operations.
AI patents granted in the United States, a paper in Nature Communications and two companies co-founded.
No jargon for those who lead, full rigour for those who build. If an idea can't be explained simply, it isn't clear yet.
No. The leadership track involves no coding: you learn what you need to decide and to use the tools with confidence. If your team is technical, it has its own track. And without tying yourself to a brand: the method matters more than the tool.
We set it during the assessment around the team's calendar. As a reference, the in-company program runs for six weeks and mentoring uses fortnightly sessions. There's no mandatory homework between sessions.
Yes. There's no fixed syllabus: the assessment defines the cases, and every session works on processes, documents and decisions from your own operation.
We only work with the information you choose to share, and cases can be anonymised. We sign a confidentiality agreement if you need one, and part of the program is knowing which data should never go into an AI tool.
Yes. We prefer to guide your team so the know-how stays in-house, but when it makes sense we build it ourselves, with AI-assisted development and human review at every step.
At the start we agree on what should change: specific decisions, processes or habits. At the close we review that list with you, with no inflated metrics.
A first conversation, at no cost and with no commitment, to understand your context and see whether working together makes sense.
We reply within one business day.