Member of Technical Staff, Quality Systems

About the role

We want someone to build the systems that automate testing for ComfyUI, which has 4 million users, 60,000+ community-built nodes, and a canvas where every workflow is unique. This is not a traditional testing role—you won’t manage manual test teams or defect queues. Instead, you’ll build the infrastructure that determines whether a release is safe to ship, enabling daily releases by proving quality through machines, not meetings.

Responsibilities

• Define what "correct" means for ComfyUI and build systems to prove it, including release gates, coverage of real user journeys, and automation to make daily shipping safe.
• Solve the combinatorial problem: design property-based and generative testing to find breakage in workflows no one has written yet, given users freely combine 60,000+ community nodes.
• Solve the visual correctness problem: build perceptual comparison tools for canvas rendering, node layout, and link routing that engineers trust, avoiding false positives from pixel diffs.
• Solve the non-determinism problem: model output shifts caused by seeds, hardware, and drivers, and build evaluation harnesses to distinguish real regressions from acceptable variance.
• Build custom node CI: verify third-party node compatibility continuously, turning our ecosystem from an untested surface into a validated strength.
• Use AI agents aggressively for exploratory testing, generated test cases, and self-repairing suites, while verifying their output to reject low-quality results.
• Make quality legible: publish metrics that track improvement and present them to the team. Later, help answer how to evaluate code quality as more of our engineering loop relies on AI assistance.

Requirements

• Built testing or verification infrastructure that other engineers depended on (this is an engineering role, not a process role).
• Worked on systems where scripted test coverage was impossible and implemented smarter solutions like property-based testing, fuzzing, simulation, generative testing, or evaluation harnesses.
• Shipped systems that gate releases and can discuss the impact when they blocked a launch.
• Strong in TypeScript or Python, and comfortable with CI systems and build graphs.
• Treat flaky tests as defects with root causes, not as "weather."
• Thrive in fast-moving environments, take ownership, and love solving problems at scale.

Nice to have

• Tested graphics, canvas, or rendering surfaces (e.g., WebGL, shaders, or visual diffing at scale).
• Built evaluation systems for non-deterministic output (e.g., golden datasets, judge calibration, scored regression tracking).
• Maintained an open-source project, especially one with a plugin or extension ecosystem requiring cross-version compatibility.
• Used ComfyUI or contributed to it.

What this role is not

• Managing a manual QA team.
• Owning a Jira defect workflow.
• Writing Playwright scripts to someone else’s specification (we already use Playwright, Vitest, Storybook, and Codecov in CI—this is the floor, not the job).

Pay

$180,000–$220,000 + equity + benefits.

About Comfy

Comfy (comfy.org) is the AI creation engine for visual professionals who demand control over every model, parameter, and output. Unlike tools that hide everything behind a prompt box, Comfy lets you connect models, processing steps, and outputs on a canvas where every decision is visible and inspectable. It empowers users to create workflows nobody’s imagined yet and share them with everyone.

What started as an open-source project in 2023 now has 4 million users, 60,000+ community-built nodes, and 150,000+ daily downloads. It’s used by artists, filmmakers, game studios, designers, researchers, and VFX houses, including teams at OpenAI, Netflix, Amazon Studios, Ubisoft, EA, and Tencent.

We’re a small, intense team in San Francisco, with backgrounds from Stability AI and Google. Many of us contributed to the ComfyUI ecosystem long before joining. We’re low-ego, high-ownership, and work hard while having fun. Everyone here is building something meaningful that will likely become our life’s work.

We recently raised $30M at a $500M valuation, led by Craft, with participation from Pace Capital, Chemistry, TruArrow, and others.