David Mashiah is an Israeli software builder who works alone.
Overview
He builds complete systems — front end, back end, data layer, deployment and the tooling that keeps them maintainable — across domains that have little in common with one another: clinical monitoring, real-time multiplayer, and local-first desktop software.
Mashiah is based in Tel Aviv, where he is a mechanical engineering undergraduate at Tel Aviv University. The software is built alongside the degree rather than out of it.
What connects the work is not a subject area but a constraint. Mashiah builds without a team, which means the usual mechanism for keeping a large codebase coherent — several people who remember how it fits together — does not exist. His response has been to move that memory out of people and into the repository itself, and it is the through-line in everything he ships.
Method
Each of Mashiah's repositories is written to be legible to an AI coding agent as readily as to a person. A generated map of the codebase, verified in continuous integration so it cannot fall out of date, lets an agent locate the right file by reading one short document rather than searching the tree. Module boundaries are enforced by lint rules rather than by convention, so a cross-feature import fails the build instead of quietly accumulating. Files are capped at a few hundred lines. A structural-debt count is permitted to move in one direction only, and the build fails if it rises.
None of this is unusual engineering practice in itself. What is unusual is applying it at the scale of one person, and treating it as the precondition for working with AI agents at all: an agent operating in a codebase with unenforced boundaries is a liability, while the same agent in a codebase that rejects invalid changes at the build step is a genuine multiplier.
Constraints
Mashiah works primarily in Hebrew and Arabic alongside English, which means right-to-left layout is a starting condition rather than a later adaptation. Several of his projects target users who cannot comfortably read small type, which sets a specification most software never has to meet: WCAG AAA contrast, a 16-point floor on body text, 48-pixel minimum touch targets, colour never used as the only signal, and a night mode treated as a first-class theme rather than an inversion filter.
His personal site implements these same rules and exposes them as live controls, so a visitor can raise the type floor, switch every low-emphasis colour to AAA, disable motion, or mirror the entire layout into Hebrew and watch the page comply.
Work
Qeben is a home-monitoring application for age-related macular degeneration, developed with advisory input from ophthalmologists and university researchers acting in a personal capacity. It pairs a daily hyperacuity self-test — designed to surface distortion in a patient's central vision between clinic visits — with a triage workflow for the clinician receiving the resulting alert. It was engineered to the standards expected of regulated software, but it was never clinically validated and never deployed, and Mashiah describes it as a research prototype rather than a product.
1 vs 100 is a live, television-format quiz game built for Tel Aviv University and deployed at cup.taueng.com, in which a single contestant faces a crowd of up to a thousand players joining from their phones. Its architecture places each game room in a single Cloudflare Durable Object, so the room's state is serialised by construction rather than by locking, and uses WebSocket Hibernation so a thousand idle connections cost nothing to hold open. The result runs roughly fifty full thousand-player games a day at no infrastructure cost. Its rules live in a pure engine with no input or output and full unit-test coverage.
Leviathan is a local-first application for study, training, habits, finance and planning, built as a static web application inside a desktop shell with no server component at all. It runs SQLite natively on the desktop and SQLite compiled to WebAssembly in the browser, using one schema and one dialect on both, so there is a single set of behaviours to reason about.
An honest scoreboard
One of these systems reached a university stage, one reached clinicians and stopped, and one runs its author's own life. None has yet become a business. Mashiah states this plainly on his own site rather than presenting the work as a track record it is not — a position he takes because in software the reasons a project stalls are usually more instructive than the reasons another one shipped, and because a claim that cannot be checked is worth less than an admission that can.