Staff Quality Engineering Lead, Gaming AI

Unity Technologies SF•Mountain View, CA

7d•Remote

About The Position

At Unity, we're shaping the future of real-time 3D technology by building AI to transform how games are created and experienced. From intelligent authoring tools to new ways of interacting with content, we're engineering the next generation of pipelines that make 3D game development faster, more immersive, and more efficient. We build AI products, and we use AI to build them. Both halves of that sentence are moving fast: new models, new harnesses, new benchmarks, and new expectations arrive continuously, and requirements change underneath us as they do. AI has made our teams dramatically quicker at designing and writing software — which means the pressure has moved downstream, onto quality. Traditional QA cannot absorb that pressure, and hiring an army of manual testers would not either. We're looking for a Staff Quality Engineering Lead to own the quality bar for the entire Gaming AI organization and to solve this the way the problem demands: with AI. You will design the systems that let us ship confidently — AI-assisted code review and remediation, automatically generated tests that stay honest as requirements change, change-aware test selection, and defect intelligence that tells the organization what actually broke and who should care. You will also take on a problem the industry has not settled: how do you judge a product whose output is non-deterministic? Our products generate code, find and fix bugs, and produce art and design. There is often no single correct answer, so correctness has to be established across repeated runs, multiple judges, and several dimensions at once. And correctness is only half of it — speed, cost, and consistency are quality attributes here, not separate concerns. A generation that is right but slow or expensive fails the creator just as surely as one that is wrong. Above all of those sits an outcome rather than an output. What ultimately matters is whether a developer can build a game faster and more cheaply than they could without us, and whether the game they build excites the people who play it. A test suite proving our models emit valid code has measured very little if creators are not shipping better games sooner. Designing the measurement chain that runs from a single generation up to that outcome is the central intellectual problem of this role, and the reason it sits at Staff level. This is a leadership role without a large team behind it. You will not test everything yourself — that does not scale, and it is not the job. You will build the processes, tooling, and methodology that raise the bar across an organization of engineers, artists, and product managers, from requirements and design through unit, feature, regression, and performance testing. You will be the gatekeeper for the quality of our AI Authoring products end to end: frontend, backend, and the upstream and downstream systems we depend on, both internal and external. Being that gatekeeper takes ownership and backbone. We move fast and we feel time-to-market pressure constantly, which means you will regularly be asked to approve something before you are comfortable with it. The job is to make the risk explicit, say no when no is the right answer even to people more senior than you, and be able to defend the call with evidence rather than instinct. It is equally the job to recognize when a risk is worth taking, say so plainly, and commit fully once the organization decides. A gate that never closes is not a gate — and one that never opens is not a partner.

Requirements

Extensive experience in software quality engineering, SDET, or quality leadership, including recent ownership of the quality bar for a multi-team organization
A track record of building quality systems — frameworks, platforms, processes, and tooling that other teams adopt — rather than primarily executing test plans
Hands-on, current use of AI and agent tooling in the software development lifecycle: AI-assisted code review, test generation, agentic workflows. We are looking for genuine daily practice, not familiarity with the concepts
Experience designing evaluation methodology for systems without a single correct answer — LLM-as-judge, rubric design, inter-rater and multi-run agreement, calibration against human judgment, and the statistics needed to tell signal from noise
Treats performance and cost as quality attributes rather than someone else's problem: latency, throughput, cost per operation, and run-to-run consistency, measured and regression-tested alongside correctness
A habit of connecting quality work to product outcomes — able to point to a quality investment that measurably changed user behavior or business results, not just a dashboard that turned green
Depth in test automation architecture across the pyramid — unit, integration, end-to-end — including CI/CD quality gates, test impact analysis and selection, and disciplined management of flaky tests
Experience building defect analytics and quality metrics that change behavior: triage and prioritization frameworks, regression attribution, and reporting that leadership actually acts on
Demonstrated ability to lead through influence rather than headcount, setting standards adopted by teams you do not manage and aligning stakeholders across engineering, art, and product
Strong ownership and backbone: a track record of holding a quality line under schedule pressure, disagreeing with senior stakeholders when the evidence warrants it, and owning the outcome either way — paired with the judgment to know which risks are worth accepting and the discipline to commit once a decision is made
Strong written and verbal communication, with the ability to make quality legible to executives and actionable for individual contributors

Nice To Haves

Game development or game QA experience, in a studio or on an engine team
Familiarity with Unity, Unreal, or a comparable real-time 3D engine
Experience validating content and asset pipelines, or evaluating generated art and design output
Performance, load, and latency testing for real-time or interactive systems
Experience standing up a quality function or practice from scratch
Familiarity with C#/.NET or Python, and with cloud backends on Azure or Google Cloud
Exposure to security, privacy, or compliance testing in a product context

Responsibilities

Own the quality strategy for Gaming AI. Define the quality bar, the standards, and the release criteria for our AI Authoring products, and hold the organization to them across frontend, backend, and dependent systems.
Build AI-native quality systems. Establish AI-assisted code review and automated remediation; drive meaningful coverage through automatically generated unit tests; generate and maintain end-to-end feature tests that track changing requirements and code; and implement change-aware test selection so that every run tests what actually matters.
Solve evaluation for non-deterministic output. Design and operate our AI-as-a-judge methodology — rubrics, multi-run sampling, multi-judge agreement, and calibration against human judgment — across generated code, bug detection and repair, art, and design.
Judge quality and performance together: correctness, latency, cost per generation, and consistency across runs are one measurement problem, not four.
Tie quality to creator outcomes. Build the measurement chain that connects an individual generation to the things that actually matter — whether creators ship faster and at lower cost with our tools, and whether the games they make land with players. Keep that chain visible enough that teams can act on it and honest enough that it survives scrutiny.
Turn defects into intelligence. Build the triage, grouping, deduplication, and prioritization systems that route developers to the most customer-impactful bugs first; attribute regressions to their cause; and give teams and leadership clear visibility into incoming, outgoing, and outstanding defects, with trends, baselines, targets, timelines, and ownership.
Shift quality upstream. Embed quality practices at the requirements and design stage, where defects are cheapest to prevent, rather than treating quality as a gate at the end of the cycle.
Drive adoption across disciplines. Partner with engineers, artists, and product managers to make the quality bar something the organization owns collectively — through tooling and process that people adopt because it helps them, not because they were told to.
Own the release decision. Hold the quality gate under real time-to-market pressure: quantify the risk of shipping, make the tradeoff explicit to the people who own the deadline, and be willing to say no and defend it. Recognize the times when the risk is worth taking, say so, and commit without relitigating.
Represent quality in technical strategy. Contribute to architectural and roadmap decisions for Unity AI, and work with owners of upstream and downstream systems, internal and external, to keep integration quality high.