top of page
ReactCon_icon.png
ReactCon_logo.png

Tomek Czajecki

Engineering Manager @ Expo
greg-fawson-foto.1024x1024 Kopie.jpg

Tomek is an Engineering Manager at Expo, where he leads the Web Services team. He has been working in the React ecosystem for close to a decade, shipped large React Native apps for several years as a mobile consultant, built backends in various technologies, and spent countless hours writing React code. He hasn’t written a full line of code in 2026, yet he has shipped more than in any year when writing code was his full-time job. He strongly believes agents can be used to ship more code and raise the quality bar if used right.

AI can write the code, but it cannot ship the app

We can now one-shot a decent app prototype with a single prompt. But it does not scale well. Codebases quickly become unmaintainable when agents keep suggesting new features that have no good product reason to exist. Task boards can fill with agent-generated activity: hypothetical bugs that were never reproduced in simulators, and features that crash the app because native APIs were used incorrectly.

This talk is about building Expo apps that get better with agents. We’ll look at how to keep agents grounded with current SDK documentation, how to use them to run real QA flows with simulator-control tooling, and how to decide which tasks to delegate to save time and which require human product taste.

bottom of page