Senior AI DevOps Engineer

Roku•Austin, TX

18h•Hybrid

About The Position

Join a high-performing, innovative team that plays a pivotal role in Roku’s mission to be the best TV streaming platform in the world. Our team is responsible for delivering intuitive, high-quality mobile applications that enhance the way millions of users interact with Roku devices globally. We pride ourselves on creating products that “just work” — seamlessly and effortlessly. This commitment to excellence is driven by a collaborative, inclusive, and results-oriented culture where your contributions will directly impact the user experience of millions. If you’re passionate about building products that feel magical and intuitive, this is the team for you. The mobile team ships the Roku Remote, Smart Home, and Howdy apps on iOS and Android. You own the CI/CD pipeline and QA automation infrastructure for the mobile engineering team. Your role will use AI to design and build a completely autonomous, self-healing CI/CD and QA automation pipeline for multiple products with millions of users. You treat AI as your primary design tool not an add-on. Every system you build should minimize human intervention, from code push to app store submission. You'll own CI Pipeline Architecture: the path from git push to a green or red signal. Your job is to make that path fast, reliable, and cheap. That includes QA Automation & Device Orchestration: the software systems that schedule, monitor, and recover the test infrastructure.

Requirements

5+ year's operating CI/CD infrastructure at scale, preferably GitLab CI
Ability to travel up to 20%
Deep understanding of mobile build systems (Xcode/xcodebuild, Gradle) and mobile-specific CI challenges (code signing, provisioning, multi-platform builds)
Strong scripting (Python, Bash) and ability to build internal tooling reservation systems, health monitors, pipeline analytics dashboards
Advanced proficiency with AI-assisted development (Copilot, Claude Code, Cursor, or equivalent) you use AI as your default approach to writing code, building systems, and solving infrastructure problems
Experience designing autonomous, self-healing systems that detect, diagnose, and recover from failures without human intervention
AI-first problem solving where your instinct is to automate with AI before adding manual process or headcount
Obsession with developer experience. You measure your success by how fast and reliably engineers get feedback, not by how complex your infrastructure is
Data-driven decision making. You measure failure rates, waste rates, device utilization, and pipeline duration and you use those numbers to prioritize your work

Nice To Haves

Experience with infrastructure-as-code (Terraform, Ansible, or equivalent) for managing cloud and on-premises infrastructure
Working knowledge of WiFi and BLE protocols enough to understand why tests that exercise radio communication behaves differently from pure software tests
Experience with mobile test automation frameworks (XCUITest, Espresso, Appium) not to write tests, but to understand what they need from infrastructure
Experience scaling CI for high-volume, automated code generation (agentic engineering, bot-authored MRs)

Responsibilities

Design and maintain CI/CD pipelines for iOS and Android on GitLab CI
Architect pipeline stages for fail-fast execution: cheapest checks first (lint, compile, static analysis), expensive checks last (device farm tests)
Build smart test routing: analyze MR diffs to determine which tests need physical devices and which can run on emulators, so 80% of MRs never touch the device farm
Build flaky test detection and quarantine systems. Classify failures as infrastructure-caused vs. code-caused so engineers trust the signal
Automate release mechanics: code signing, versioning, TestFlight/Play Console uploads, dSYM and mapping file management. The goal is zero manual steps between merge and app store submission
As agent-authored MR volume grows, ensure pipelines absorb the increase without degrading speed or starving human-authored MRs of resources
Build the device reservation and orchestration system that assigns devices to CI jobs, prevents contention, and maximizes utilization without manual scheduling
Design self-healing automation: health checks detect unresponsive devices, trigger remote recovery via API, and re-register them no human intervention required
Define the device compatibility matrix which firmware/model combinations require real hardware, and which can run on emulators
Implement priority-based test routing: device-touching MRs get farm time, UI-only MRs never queue for a device
Use AI to identify failure patterns, predict infrastructure issues, and continuously optimize pipeline performance