Architect and external R&D partner for architecture and engineering practices.
I'm an architect who built the tool he wished existed. I help practices use climate data, computation and AI on real projects, without the hype: I investigate a hard question, prototype it, and show you what is actually possible.
Which one sounds like you?
I want to learn to build my own agent.
A live class, then a cohort: a small group over four weeks.
▶ LearnI have one task I want automated.
I build it for you: a simple agent or an agent app.
▶ BuildMy firm needs to work out where AI, climate and computation fit.
Mapping session, R&D Kick-off, R&D Partnership.
▶ For firmsAn independent R&D partner for architecture & engineering.
You have an interesting problem. It sits somewhere between climate, design and AI, and no one on the team has time to chase it. I investigate it, prototype it, and show you what is actually possible.
Short, bounded investigations that end in something working, not a slide deck.
Finding the specific tasks where AI and agents improve design and business workflows, and building the proof.
Translating changing climate conditions into practical implications for sites, assets and projects.
Daylight, comfort and performance thinking turned into computational workflows your team can run.
※ NOTEBased in Lisbon. I work on site with practices in Paris and Oslo, and online with practices anywhere else. A small number of clients, deep work, alongside building Spacio.
A hands-on live class for architects, engineers and designers, and anyone else who repeats the same task every week. You leave with a small agent that does one real task from your own practice, and a clear sense of where agents help and where they don't. Afterwards, a small group continues as a four-week cohort.
Most environmental analysis is done correctly and too late. What's missing is the translation between metrics and design moves.
Retrofit is the right answer. New build will keep happening anyway. Both are true, and that tension is the job.
Iteration speed, environmental and code complexity, and renders you can argue over. Specific tasks, not a revolution.
