I'm Malte Risto — experimental psychologist and UX researcher. For fifteen years I've worked one question: how do humans and intelligent agents coordinate?
Today I build open practice infrastructure for human–AI collaboration, and I help teams learn to use AI on their work — under their privacy and tooling constraints — not as generic prompt training.
Facilitation & workshops. Hands-on sessions where people practice with AI on real tasks and real documents (grant review, research workflows, knowledge work). I design for the tools they are allowed to use — including institution-hosted models and RAG — and for the failure modes that show up in practice: wrong model shelf, brittle retrieval, hallucinations, “it worked in the demo.”
Practice infrastructure. Magic is an open framework for intentional human–AI partnership. turtleOS is a place for practices — rooms with a Turtle and a River, on a machine you own. I build what I use with my own family and practice.
Research stance. I come from studying trust, advice-following, and coordination with automated systems — then designing those interactions in industry — not from selling AI products. The unit of interest is the human–AI pair doing real work.
If you are choosing between a platform/service-center training offer and an external facilitator: platforms teach the rails; I help your people become competent on those rails for the jobs they already have.
PhD, Experimental Psychology — University of Twente & TNO (2014). How drivers respond to, trust, and integrate AI-generated tactical advice (Cooperative In-Vehicle Advice).
Postdoc — UC San Diego Design Lab (2015–2017). Human interaction with automated vehicles with Toyota and Nissan R&D; Wizard-of-Oz studies of pedestrians and seemingly driverless cars on real roads. ACM UbiComp '16, Driving Assessment '17.
UX Researcher — Volkswagen Group Future Center Europe, Potsdam (2018–2024). Early-stage human–vehicle interaction for autonomous and future vehicles, in a multinational design/research team. The center closed at the end of 2024.
Independent (2025–). Building Magic and turtleOS; writing in public about cognitive sovereignty and human–AI partnership; exploring how this work can sustain a family without employment as the default path.
Machine-readable detail: career.md. Scholar: Google Scholar.
- Task-first. Start from the work (e.g. checking an application against funding rules), then the tool settings that make that possible.
- Privacy-real. Sensitive work stays on approved infrastructure; I don't push consumer chatbots where policy forbids them.
- Honest about limits. Retrieval systems fail in characteristic ways; I teach people to see those failures instead of performing confidence.
- No cult, no pitch deck. I share practice; I don't recruit. More on that boundary: older note in stance.md and the book.
Based in Berlin. Languages: German (native), English and Dutch (full professional).
malte.risto@gmail.com · X (@malteristo) · this repository
For what I'm focused on right now: now.md. For how I think more broadly: stance.md.
—Malte Risto