Uploaded September 2025 | Updated September 2026, 1 hour ago
At HelloFresh, migrating existing iOS and Android apps to React Native meant rethinking how to deliver consistent cross-platform UI at scale.
In this talk from React Universe Conf 2025, Ohans Emmanuel (Staff Software Engineer at HelloFresh) shares a real-world journey of using GenAI coding agents to automate and accelerate the conversion of web components into native mobile UI.
Skip the intro ⏩ 0:39
🔹 The challenge of migrating a massive codebase with 10,000+ React components, including a critical Product Card with 13k lines of code.
🔹 Why naive “one-shot” AI conversions fail, and how to work around large contexts, complex states, and deterministic requirements.
🔹 Techniques like selective context, breaking work into smaller agent tasks, and enforcing deterministic states with TypeScript.
🔹 Building self-correcting agentic flows: having agents write failing tests first, validate implementations, and iterate until outputs matched expected APIs and UI.
🔹 How these methods reduced weeks of manual migration work to just hours, while improving consistency across platforms.
A practical, battle-tested case study of how GenAI can meaningfully accelerate cross-platform React Native adoption.
See you at another React Universe Conf 🎟️ clstk.com/3JP8NEB
Follow React Universe Conf on X to stay up to date 🐦 https://x.com/reactuniverse_
Check out Callstack content 📚 clstk.com/4mbeni5
At HelloFresh, migrating existing iOS and Android apps to React Native meant rethinking how to deliver consistent cross-platform UI at scale.
In this talk from React Universe Conf 2025, Ohans Emmanuel (Staff Software Engineer at HelloFresh) shares a real-world journey of using GenAI coding agents to automate and accelerate the conversion of web components into native mobile UI.
Skip the intro ⏩ 0:39
🔹 The challenge of migrating a massive codebase with 10,000+ React components, including a critical Product Card with 13k lines of code.
🔹 Why naive “one-shot” AI conversions fail, and how to work around large contexts, complex states, and deterministic requirements.
🔹 Techniques like selective context, breaking work into smaller agent tasks, and enforcing deterministic states with TypeScript.
🔹 Building self-correcting agentic flows: having agents write failing tests first, validate implementations, and iterate until outputs matched expected APIs and UI.
🔹 How these methods reduced weeks of manual migration work to just hours, while improving consistency across platforms.
A practical, battle-tested case study of how GenAI can meaningfully accelerate cross-platform React Native adoption.
See you at another React Universe Conf 🎟️ clstk.com/3JP8NEB
Follow React Universe Conf on X to stay up to date 🐦 https://x.com/reactuniverse_
Check out Callstack content 📚 clstk.com/4mbeni5
