Seven things I'm regularly asked for. Most come from programmes, departments and research groups, and most start with an email describing a cohort, a module or a deadline.
Rates depend on the format and the institution. Tell me what you need and I'll come back with a number and an honest view of whether I'm the right person, which sometimes I'm not.
One-to-one or small-group reviews for students working with living materials, biodesign, or AI-driven biology. Undergraduate through doctoral.
I read the work on its own terms, then ask the questions an external examiner or a funder would. What did you actually make. What did you learn that you couldn't have predicted. What would you have to show someone to make the claim stick.
I've supervised MSc students at TU Delft, written the assessment framework for a 7-day intensive with ten project groups at Shih Chien, and judged for iGEM and the Biodesign Challenge.
Remote or in person. Single sessions, or a block across a term.
Keynotes, conference talks and panels on biodesign, living materials, DNA data storage, microbe-HCI, and what generative models are doing to biology.
Recent: the biodesign panel at Flanders Technology & Innovation in Brussels, October 2026. KTH Stockholm, October 2025. Next Nature Conference, Eindhoven, October 2025. Earlier: the day-two opening keynote at Renewable Futures, and BioFabbing Convergence at CERN.
I speak at universities without a fee. Festivals and commercial events pay.
Guest lectures, seminar series and research-group talks, built for the cohort rather than delivered off a shelf. I've done these at Politecnico di Milano, KAIST, Feng Chia, Shih Chien, Hertfordshire and the RCA.
Format: a single 90-minute lecture with discussion, a two-session pair spread across a week, or a doctoral seminar for a small research group. Remote works.
Hands-on sessions where people work with the material or with the model rather than hearing about it. I've run these from 3 hours to 7 days.
The ready-made one is AlphaFold Futures, three days on what structure prediction does to design practice, first run in July 2025 for an international cohort.
Everything else is built to the brief. Methods I use: futures cone and Dator's four futures, STEEPI signal scanning, backcasting, card-based ideation. Remote delivery works, and I've run it.
For departments building or rebuilding a programme where biology, design and AI meet. The question is what students should be able to judge by the time they graduate, and whether the current structure gets them there.
The most recent was for a US art and design college: a written review built on semi-structured interviews and comparative desk research, on the judgment capacity students need to evaluate AI-generated biological proposals.
I also own a six-week AI × Biodesign curriculum with three original teaching models, a six-course biodesign co-major architecture, three full syllabi, nineteen numbered toolkits, and two further courses written for Master's and Bachelor's cohorts. So I'm reviewing from having built the thing rather than from a framework.
Scoped per programme.
Explainers, technical reports, workshop and project reports, and editing for research groups, institutes and programmes whose science is hard to write about. Protein design, biofabricated materials, DNA data storage, synthetic biology, and the AI layer on top of them.
I can read the pipeline and the paper, so you're not paying me to interview someone who can. My own reports are on Zenodo and my peer-reviewed work is mostly at ACM venues, so the register is checkable before you commission.
A reader for a draft before it goes in. Structure, the claim the reviewer is actually assessing, the gap between what you did and what the text says you did, and whether the methods section survives a sceptical read.
I'm a registered expert on the European Commission's Funding & Tenders database, I review for ACM conferences and design journals, and I have written and won a competitive national award myself: 10,000 GPU-hours on the UK's national AI research resource, as sole applicant.
Best used two to three weeks before the deadline, while there's still time to change something.