Expo spotlights Shadosen, built by agent team
Expo featured indie developer Yugo Ogura's language-learning app Shadosen as an Expo App Awards showcase, spotlighting an end-to-end multi-agent workflow where AI agents handle product, design, QA, and coding while the developer approves pull requests. The setup pairs autonomous agent implementation with Expo Updates to deploy continuous over-the-air fixes directly to users.
Shadosen offers a compelling real-world preview of the solo-dev agent team stack, where an engineer acts as executive editor and reviewer rather than syntax author. Coupling autonomous multi-agent dev loops with instant over-the-air updates removes the release friction that typically bogs down mobile developers.
- –Multi-agent specialization: Dividing tasks among dedicated AI agents for product management, interface design, QA testing, and code generation turns solo development into an orchestration and PR review discipline.
- –Rapid iteration via Expo Update: Shipping agent-written fixes via over-the-air updates bypasses App Store review latency and lets bug fixes reach users within minutes.
- –Production-grade proof of concept: Demonstrates that autonomous agents can build and maintain a complete production iOS app with complex audio processing, dictation tracking, and AI pronunciation scoring.
- –Human-in-the-loop governance: The developer retains absolute veto power over every merge and release, keeping agent hallucinations and architectural drift in check.
DISCOVERED
10h ago
2026-11-16
PUBLISHED
10h ago
2026-11-16
RELEVANCE
AUTHOR
expo