Designer.
Data practitioner.
Builder.
I work between three practices: product design, data analytics, and experimentation.
Selected work · 2025 — 2026
4 of 4
On Connecting Tokens to Embeddings
Tracing and visualising the encoding of human-readable text into embeddings that are consumable by a neural network.
On Spaced Repetition as a Knowledge Management System
How a marriage between the right algorithm and data structure can unlock multiple possibilities in a knowledge management system.
Brain Reps: Building a Reproductive Endocrinology Research Tutor
A graph-augmented spaced-repetition research agent built solo in 10 days — Neo4j, Qdrant, Cloudflare Workers, and Ollama, at $0 cloud cost. From monolith to multi-channel production.
Improving Access to Fertility Education & Medical Management Services
A booking and landing page for a FEMM-certified physician. Designed to convert a curious visitor into a confirmed appointment in two clicks. Built with a split-panel layout, interactive Physarum canvas, and content strategy precise enough to handle reproductive health with both clinical rigour and human warmth.
Refinement is the core and no version is final.
Client work and personal projects that follow the principles of human-centered design. Every project is a live effort to improve upon the observations and feedback of the target customer.
Enter Metamorphosis →
Knowing as a working method.
Expositions of insights from inside data and AI work. Case studies and reflections on what was tractable and what wasn't.
Enter Gnosis →A working notebook of experiments.
Active, paused, complete, and abandoned experiments. All four states stay visible. This is where the next idea is born.
Enter Tinker →
One person, three practices, kept in public.
I'm a product designer, developer and AI/data practitioner based in Nairobi. I learn by building. Half my work is for clients - usually small teams trying to do something new, while the other half are my own experiments, and made public on this site.
I work across design, software and applied data science. I try to keep the different practices in conversation with each other — to let experiments inform what I do for clients, and let client work raise questions that pull me into new territory. The work is mostly for clients, but I write about it on this site because I find it more useful to see how things came together over time than to see the final shape.
Read the long version →