Senior AI DevOps Engineer

Roku•San Jose, CA

8h•$141,300 - $305,700•Hybrid

About The Position

Roku is seeking a Senior AI DevOps Engineer to own CI/CD pipelines and QA automation infrastructure for its mobile engineering team, which ships the Roku Remote, Smart Home, and Howdy apps on iOS and Android. The role will use AI as a primary design tool to build autonomous, self-healing CI/CD and QA automation systems spanning multiple products. The goal is to minimize human intervention from code push through app store submission. The engineer will own CI pipeline architecture, optimizing the path from Git push to a reliable pass/fail signal, and will build systems to schedule, monitor, and recover mobile test infrastructure. The work includes improving pipeline speed, reliability, and cost; orchestrating physical devices and emulators; automating releases; and using AI and data to detect failure patterns and continuously improve performance.

Requirements

At least 5 years of experience operating CI/CD infrastructure at scale, preferably with GitLab CI.
Ability to travel up to 20%.
Deep understanding of mobile build systems, including Xcode/xcodebuild and Gradle, and mobile-specific CI challenges such as code signing, provisioning, and multi-platform builds.
Strong scripting skills in Python and Bash, with the ability to build internal tools such as reservation systems, health monitors, and pipeline analytics dashboards.
Advanced proficiency with AI-assisted development tools such as Copilot, Claude Code, Cursor, or equivalent, and use of AI as a 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 approach, with an instinct to automate using AI before adding manual processes or headcount.
Strong focus on developer experience, measuring success by how quickly and reliably engineers receive feedback.
Data-driven decision-making, including measuring failure rates, waste rates, device utilization, and pipeline duration to prioritize work.

Nice To Haves

Experience with infrastructure-as-code tools such as Terraform, Ansible, or equivalent for managing cloud and on-premises infrastructure.
Working knowledge of WiFi and BLE protocols sufficient to understand how radio-communication tests differ from software-only tests.
Experience with mobile test automation frameworks such as XCUITest, Espresso, or Appium, with an understanding of infrastructure needs.
Experience scaling CI for high-volume automated code generation, including agentic engineering or bot-authored merge requests.

Responsibilities

Design and maintain iOS and Android CI/CD pipelines using GitLab CI.
Architect pipeline stages to run inexpensive checks such as linting, compilation, and static analysis before more expensive device-farm tests.
Build smart test routing that analyzes merge-request changes to decide which tests require physical devices and which can run on emulators.
Build flaky-test detection and quarantine systems, distinguishing infrastructure-caused failures from code-caused failures.
Automate release mechanics, including code signing, versioning, TestFlight and Play Console uploads, and dSYM and mapping file management, aiming to eliminate manual steps between merge and app store submission.
Scale pipelines to handle increasing agent-authored merge-request volume without slowing pipelines or depriving human-authored merge requests of resources.
Build device reservation and orchestration systems that assign devices to CI jobs, prevent contention, and maximize utilization without manual scheduling.
Design self-healing automation that detects unresponsive devices, initiates remote recovery through APIs, and re-registers devices without human intervention.
Define the device compatibility matrix, identifying firmware and model combinations that require real hardware versus emulators.
Implement priority-based test routing so device-dependent merge requests receive device-farm time and UI-only merge requests do not wait for a device.
Use AI to identify failure patterns, predict infrastructure issues, and continuously optimize pipeline performance.