Start with the customerTechnical judgment begins with the user’s problem. Elegance matters when it produces a meaningfully better outcome.
Build the systems that let AI operate software
Join the founding team building the runtime where agents build, run and verify every change on cloud devices and Macs.
- 4 open roles
- San Francisco, CA
From the founding team
The hard part is understanding, operating and trusting software
Revyl is building the runtime for that next loop, where people and agents build real apps and verify every change with evidence.
We’re a small San Francisco team. We built DragonCrawl at Uber AI and are backed by Felicis, General Catalyst, and Y Combinator. Come build with us.
The Revyl Team, San Francisco, CA
Open · August 2026
We’re hiring now
- Staff Infrastructure Engineer — Device CloudOwn the technical strategy and systems that scale the device cloud and distributed compute platform behind Revyl.
- Team
- Engineering, Infrastructure
- Location
- San Francisco, CA, US · Remote (US)
- Type
- Full-time
- Software EngineerBuild the AI agent loop, device infrastructure, workflow orchestration, and product surfaces of the platform.
- Team
- Engineering, Full stack
- Location
- San Francisco, CA, US
- Type
- Full-time
- Engineering InternShip production work across agent systems, device and Mac infrastructure, and product with high ownership alongside the founders.
- Team
- Engineering, Full stack
- Location
- San Francisco, CA, US
- Type
- Internship
- GTM InternCreate content, talk to customers, and ship projects that grow the business alongside the founders.
- Team
- Growth and marketing
- Location
- San Francisco, CA, US
- Type
- Internship
How we work
The team we’re building
Own ambiguous problemsYou make decisions with incomplete information, communicate the tradeoffs, and carry the work from idea to production.
Work across the systemThe best answer may cross agents, infrastructure, device runtimes, developer tooling, telemetry, and product experience.
Experience that stands out
Each of these can help you reach the hard parts faster. None is required.
- Distributed systems at scale
- Device and compute fleet infrastructure
- LLMs, AI infrastructure, or agent systems
- Workflow orchestration
- OpenTelemetry or distributed tracing
- CLI or developer tooling
Probably not a fit if
You want the edges defined for you.
- You want work broken into well-scoped tickets.
- You are uncomfortable making architectural decisions before every variable is known.
- You prefer staying inside one narrow layer instead of following a problem across the stack.
The work
Problems worth owning
The AI agent loopMake capture, grounding, execution, and reflection reliable as agents observe screens and operate real applications.
Workflow orchestrationCoordinate test execution across distributed pools of devices, services, queues, retries, and long-running workflows.
Device cloud infrastructureBuild the systems behind Android emulators, iOS simulators, real devices, streaming, automation, and isolation.
The developer loopConnect local development to cloud verification through the CLI, build uploads, live sessions, and CI integrations.
Observable executionsUse OpenTelemetry, trace correlation, and purpose-built debugging tools to explain what happened across the system.
Real-time product surfacesShape WebSocket streaming, live device views, graph-based test visualization, and the web, CLI, and CI experience.
How to apply
We’d love to hear how you think
Tell us which role interests you, share your answer, and include links to work you are proud of.
What do you think the future of software development looks like, and what are most people missing about it?
- 02Show us how you buildShare a project, problem, or artifact that shows how you think, learn, and turn ambiguity into something real.
- 03Start the conversationApply on our Y Combinator page with the role, your answer, and links to work you are proud of.
Build the runtime for cloud agents
Apply on Y Combinator with the role you’re interested in, your answer to the question above, and links to your work.