Micromobility platformBuilding
Flash Bikes
The problem: shared e-scooter platforms are usually imported wholesale into the Gulf, then retrofitted for local rules, language, and payments. The seams show everywhere: in compliance, in Arabic layouts, in rider trust.
The thesis: build the platform for Dubai from the first commit. RTA permit conditions as product requirements, Arabic RTL as a first-class experience, AED-native payments, and a server-authoritative fleet model designed for real streets. Rider app, operator platform, and fleet backend are all in active build.
Live sports engagementPrototype
Playdaq
The problem: watching live sport on a second screen is passive. The momentum everyone feels in a match has no visual, interactive form.
The thesis: render match momentum as a living probability curve, with a locked design foundation that keeps the product's identity stable no matter which teams play. Interface prototypes and the full design system exist; the engine decides next whether it graduates to build.
Engineering infrastructureInternal R&D
Rivora Fleet
The problem: a small team cannot ship venture-grade software on multiple fronts with conventional processes. Speed usually costs discipline.
The thesis: coordinate specialised AI engineering agents, each with a role, a doctrine of hard-won engineering lessons, and non-negotiable review gates, under a human founder's final pen. The Fleet built and maintains Snook A Look in production and is now building Flash Bikes. It is infrastructure, not a product for sale, and it is the engine's biggest unfair advantage.
Developer toolingExploration
Graphify
The problem: understanding a large unfamiliar codebase or document set is slow, and the map you build in your head disappears when you close the editor.
The thesis: generate a navigable knowledge graph from any input: code, docs, papers. Communities, hub nodes, and an explorable wiki instead of a folder tree. It already runs inside our own projects; whether it becomes a public tool is an open question the lab is testing.
Concepts stay in the lab until they earn a status. When a venture is archived, it will say so here. What you see is the honest state of the engine.
