Automated CI runs in twelve months
GitHub Actions audit04 defects detectedBUILD FAILED
01/05QA Automation Lead · South Jordan, Utah
Something's broken.Some-thing'sbroken.
I'll find it.
I'm Brayden Pay. I lead QA at Fluid, write Playwright tests, train testers, and ask the annoying question before it becomes an expensive one.
Currently: QA Automation Lead at Fluid
Open to the right problem ↗Measured in the work
22-month source audit · Sep 2026Four signals, one pattern: quality systems that increase coverage, catch risk earlier, and give teams a clearer release decision.
Regressions caught before release in 2026
2026 issue-label auditGrowth in automated test coverage
Repository history auditPeople led across QA and shipping
Largest team scopeFrom evidence to value
What a new team getsBuild the system
6 → 61 test specs and 51 CI workflows show scattered checks becoming infrastructure a team can rely on.
Result: broader coverage without matching headcount growth.Make the signal useful
200+ regressions caught before release matter because every failure is paired with evidence and a clear verdict.
Result: faster decisions and fewer customer-discovered defects.Test the risky parts
Payments, native iOS, Electron, and AI systems require different tools, but the same standard of proof.
Result: one quality lead who can follow risk across the stack.Raise the team
Training, work assignment, and review turn individual knowledge into repeatable team practice.
Result: testers become more independent and releases become calmer.Proof in motion
Audited growth · Not estimatesAutomation platformOct 2025 → Sep 2026
Test specsCI workflows
Merged pull requests authored
768 of 972 · Repository auditThe tests get tested, too
74Harness tests protecting the automation platform itselfProof that reliability applies to the quality system—not only the product.What automation buys
Assumptions shown · Source audited208 cases at four minutes each. CI runs the same coverage in minutes, without taking a tester away from exploratory work.
Hourly staging checks plus daily production checks each year—coverage a manual schedule could never realistically reach.
Three verdicts, not two
Accuracy before certaintyA binary system turns blocked work into a lie. The QA system preserves a third answer when the evidence is not strong enough.
Behavior verified live against the acceptance criteria, with evidence attached.
A real defect, reproduced and documented with an honest status trail.
Blocked, not deployed, or not safely testable. Never forced into a false verdict.
QA that can move the fix forward
Sometimes the fastest bug report is a pull request.
When a problem is contained and low-risk, I use agent-assisted development to trace the cause, draft the change, validate the behavior, and push a review-ready PR. Engineering gets evidence and a credible fix—not another ticket waiting in the queue.
What this provesEnough technical range to shorten the handoff, with human judgment owning the scope and verification.Beyond the browser
Three toolchains · One standardWeb
PlaywrightEnd-to-end coverage across portals, checkout, storefronts, and admin surfaces.
Native iOS
MaestroThe real app binary on a simulator—not a mobile-sized browser pretending to be one.
Electron
Raw CDPDesktop flows driven directly through the Chrome DevTools Protocol.
Why payments belongs here
Domain depth · Not a buzzwordA checkout can succeed and still be wrong.
Follow where the money went—not only whether it moved.
Payment routing has silent failure modes: the wrong gateway can approve a charge, an inactive rule can fall through incorrectly, or a claimed transaction source can bypass the intended bucket. Those cases demand an oracle that verifies processor identity after settlement.
What this provesHigh-risk domains can be learned deeply enough to design the test strategy—not just execute somebody else's checklist.
Independent product / 01
DIGIMORA
A number-memory training game built end-to-end across two fully separate native codebases: SwiftUI for iOS and Kotlin with Jetpack Compose for Android.
APPLESubmitted · In reviewGOOGLE PLAYClosed testing in progress
Verification stayed part of the build.
I traced a cross-screen dark-mode failure to SwiftUI environment scope, tested timing races and relaunch persistence, checked both themes on both platforms, and caught signing-key and build-artifact risks before they reached version control.
What it proves: root-cause discipline across native products.Two releases. One moving plan.
I managed signing, store compliance, pricing, listings, banking, device verification, and feature parity across two different release systems. When Google's device step blocked launch work, I moved the store listing forward instead of letting the project stall.
What it proves: dependency-aware delivery without a project manager standing overhead.The failures that stay quiet
Five systems · Same instinctThe most expensive failures rarely announce themselves.
A payment succeeds on the wrong gateway.A point tap lands on empty space and passes.A retry setting is accepted but never runs.A citation is correct while the answer is wrong.A test turns red because the documentation improved.
Selected systems
Quality · Automation · OperationsTest architecture · Case study coming soon
Automation Platform
A Playwright system that grew from six specs to 61, with 208 cases, 51 active workflows, and 74 tests protecting the test harness itself.Workflow automation · Case study coming soon
QA Operations Bot
A Slack system that answers coverage questions, runs live checks, gathers useful bug context, and routes the result without pretending a guess is a verdict.Agent engineering · Case study coming soon
Agent Knowledge System
A maintainable knowledge layer for AI-assisted QA: grounded selectors, versioned prompts, cited references, a critic gate, and an honest “inconclusive” state.Independent product · Store review
Digimora
A 60-level number-memory game I designed, built, tested, and submitted solo across separate native iOS and Android codebases.What I can do
Capability over companyTest architecture
Reliable automation starts around the tests: fixtures, selectors, suite boundaries, failure evidence, and checks for the harness itself.
Release systems
Repeatable checks become CI that runs without babysitting and tells the team what failed before a customer does.
Agent engineering
Agents stay grounded in real source material, their misses get measured, and human judgment remains between a lead and a ruling.
Payments QA
Coverage follows the money through routing, transaction types, fallbacks, 3DS, webhooks, and the case where a charge succeeds through the wrong processor.
More than browsers
The toolchain spans Playwright for web, Maestro for native iOS, and the DevTools Protocol for Electron desktop apps.
Team enablement
Training, review, and durable documentation turn knowledge in one person’s head into a system the whole team can use.
Experience
2020—PresentFluid
QA Automation Lead
Promoted to lead after joining the QA team. I’ve managed as many as 10 people across QA and shipping, trained new testers, assigned their work, and built Playwright coverage around the releases that matter.
Tesla
Data Analyst
Worked through large vehicle datasets, checked questionable records, and brought the problems I found to engineering and operations.
Wasatch Property Management
Financial Manager
Handled the paperwork, vendors, and day-to-day financial work for two senior communities. When something stalled, I tracked down what was missing.
City of Eagle Mountain
Customer Relations Representative
Helped residents with billing and city services, updated web content, and supported public events. It taught me how quickly a small system problem becomes somebody’s whole afternoon.
How I work
- 01Detect
- 02Isolate
- 03Prove
- 04Release
Break it.
Bring receipts.
I turn failures into decisions. Every report defines the impact, the reproduction path, the supporting evidence, and the condition required for release.
I keep releases calm, develop strong testers, and align product and engineering around the same facts.
PlaywrightTypeScriptUI AutomationEnd-to-End TestingRegression TestingExploratory TestingGitHubLinearSlack AutomationVercelMaestroPower Automate
Education
Bachelor of Science, Strategic CommunicationUniversity of Utah · 2023