**What you'd work on**
\- The AI agent loop. Our agents observe device screens, ground actions to UI elements, execute steps, and reflect on whether they worked. It's a capture-ground-execute-reflect cycle and there are a lot of hard problems in making it reliable
\- Distributed workflow orchestration for test execution across pools of devices
\- Device infrastructure for Android and iOS. Emulators, real devices, streaming, automation
\- Our CLI that ties local dev to cloud verification. Build uploads, live device sessions, CI integration
\- OpenTelemetry integration, trace correlation, and the tooling we use to debug executions
\- Real-time execution UX. WebSocket streaming, live device views, graph-based test visualization
\- Building out product surface (CLI, web app, CI integrations)
**What we're looking for**
\- High customer empathy. You think about the user's problem when you're designing a technical solution, not just the technical elegance
\- You enjoy novel problems with real technical and product ambiguity
\- You've built and scaled systems that run in production under real load
\- Practical experience with LLMs, AI infra, or agent-based systems
\- Track record of high ownership and autonomous execution
\- Strong communication
**Experience that would stand out**
\- Mobile build tooling experience (Swift, Kotlin, Flutter)
\- Mobile testing infrastructure, device automation, or mobile platform tooling
\- OpenTelemetry or distributed tracing
\- Workflow orchestration systems
\- CLI or developer tooling
**Probably not a fit if**
\- You want well-scoped tickets
\- You're uncomfortable making architectural decisions without full information
\- You prefer deep specialization in one layer over working across a full stack
Revyl is building AI-native mobile infrastructure. We automate software reliability, starting with end-to-end testing that lets companies ship faster with confidence. You write tests in natural language, our AI agents execute them on iOS and Android devices, and you get back full execution traces. The underlying primitive is AI agents that can actually operate phones. Testing is the first application.
The founding team built DragonCrawl at Uber AI and left to commercialize it. We're backed by Felicis, General Catalyst, and Y Combinator, with angels including the co-creator of OpenTelemetry and people from Meta, Nvidia, and Uber. We have early enterprise traction and are building toward being the default reliability platform.
**Why join?**
\- Hard technical problems (AI agents, distributed systems, mobile device infrastructure) not CRUD apps
\- Shape technical and product direction from the ground floor
\- Flat hierarchy, high ownership, full autonomy
\- Small, talent-dense founding team in San Francisco
\- Competitive compensation and equity
Visa sponsorship is not available for this role.
