Malik Chohra

I built and published an open-source generative UI SDK for React Native (v0.1.3), featuring 11 built-in components with Zod-validated schemas and 5 LLM adapters, including OpenAI, Anthropic, Gemini, Ollama, and Google's A2A protocol, enabling an LLM to ship native UI to a mobile app turn by turn.

I engineered a React Native boilerplate designed to eliminate the six weeks of repetitive setup every AI mobile app requires. It comes in three tiers: a core stack of Expo, Supabase, Redux Toolkit, RTK Query, Restyle, and RevenueCat, plus a dedicated AI tier with both local and cloud AI integration and generative UI support.

ScorePlay is the media management platform for the sports industry. 200+ rights holders, leagues, and teams use it (UEFA, Premier League, NBA, NHL, Paris 2024) to organize, tag, and distribute their photo and video content.

The company is Series A ($20 million), backed by Alexis Ohanian, Kevin Durant, and Giannis Antetokounmpo.

I led mobile for the React Native (Expo) app, the surface athletes, media teams, and clubs use to grab and share media from wherever they are. I ran a team of 3 engineers and worked directly with the CTO, CEO, and mobile product manager.

The app is media-heavy, so performance was a constant fight. I wrote native modules in Kotlin and Swift for background fetching and bridged them into React Native, then used Datadog to find what was slow and fix it.

Features I shipped include search, authentication, and deep linking.

I also ran an AI-assisted, spec-driven development pilot for the mobile team. After that, I created the entire AI assistant workflow for our mobile teams using Claude Code, and we are now taking that approach to the front end.

DocMorris is a German digital pharmacy serving 9+ million users. I led the mobile team of 6 through one of its most trust-critical projects: integrating Germany's NFC electronic health cards as a sign-in factor for the app.

That meant native module work on both platforms, Swift on iOS and Kotlin on Android. Android was the hard part. NFC chipsets behave differently across phone models, so we hit fragmentation bugs deep into testing and still shipped on time by cutting non-essential features from v1.

I also pushed the app's quality numbers. Crash-free rate climbed to 96%, and the performance index went from 60 to 80, through observability with Datadog, a FlatList migration for heavy lists, and code splitting to shrink the bundle.

The last piece was a backend-for-frontend layer I built in NestJS. Mobile clients were hitting 7+ services directly, which was slow over mobile networks. The BFF handled projection and batching, cut payloads roughly 4x, and went live without an outage.