Arthur Locke
Senior QA Engineer
Senior QA Engineer with 8+ years of experience across manual and automated testing on PC, mobile, console, and cloud-native platforms. The last 1-2 years have centered on building test automation frameworks, automated integration test suites, and AI-augmented QA tooling — hands-on with Playwright, API integration testing, and CI/CD-integrated continuous testing. Proven track record of leading QA practices and driving AI adoption across engineering teams.
Projects
- Designed and built a multi-package platform that migrates manual TestRail regression suites into AI-evaluated, executable regression tests against a real Unreal Engine 5.8 target — built end-to-end through 12 planned milestones, each shipped with working code, passing tests, and documentation.
- Deterministic pipeline: normalizes raw manual TestRail cases, evaluates each against live-discovered engine capabilities to decide automatability, generates a versioned YAML regression spec, gates publishing behind an explicit human-approval step — run for real against 9 seeded cases, correctly automating 7 and rejecting 2 for a proven capability gap.
- Custom MCP server (testrail-mcp) wraps TestRail's REST API v2 with config-driven, per-project read/write/delete scoping and a soft-delete preview mode, verified live including the write-rejected/write-allowed/soft-preview/real-delete round trip.
PythonPydanticuv workspacemcp SDKSQLiteDocker
- Designed and built a self-hosted AI support-ticket triage and classification platform on top of Chatwoot (unmodified), architected as independently swappable pieces — two self-hosted MCP servers and Claude as a replaceable intelligence provider — documented across 6 ADRs recording the actual reasoning, not a stated principle.
- Found and fixed 10 real cross-service bugs by running the full stack against a live Chatwoot instance and real Docker containers rather than trusting unit tests alone: Chatwoot's SSRF protection silently blocking same-network webhook delivery, a labels API with replace-not-merge semantics, an inconsistent JSON envelope between two Chatwoot search endpoints, and an asyncpg connection pool silently broken by spanning two different asyncio event loops.
- 92 automated tests (pytest, moto for S3 mocking) across four services, all passing; every feature verified live against a running Docker Compose stack, not left as untested documentation.
PythonFastAPIChatwootClaude APIPostgreSQL (pgvector)Docker ComposeGoogle Cloud Run
- Built an enterprise multi-agent AI platform across the full Microsoft stack, verifying every architectural capability live rather than trusting documentation — including catching and documenting a genuine platform-side OAuth consent bug in the Outlook connector across three separate investigation attempts, and confirming two Dataverse/Copilot Studio ALM limitations against Microsoft's own documented known issues.
- Two independent CI/CD pipelines in GitHub Actions, each verified end-to-end via live deployments rather than assumed correct: one redeploys the entire Azure infrastructure from Bicep, the other promotes both business data and the multi-agent Copilot Studio deployment itself between environments.
- Root-caused and fixed real, non-obvious issues via live server responses and direct API queries: a GitHub OIDC subject-claim format change, several azd-in-CI environment-state gaps, and a tenant security policy silently blocking an OAuth login flow.
AzureBicepCopilot StudioPower PlatformSemantic KernelEntra IDGitHub Actions
- Designed and built a self-hosted observability platform ingesting real, native OpenTelemetry telemetry from AI coding agents (Claude Code, Codex CLI, Gemini CLI) — deliberately no synthetic data generator — normalizing each vendor's distinct schema into one shared ingestion pipeline backed by SQLite.
- Verified end-to-end with a real, live-connected Claude Code session flowing through the OpenTelemetry Collector (OTLP over gRPC/HTTP) into ingestion and the Next.js dashboard — not a test payload.
- GitHub Actions CI gate on every push (lint/typecheck/test/build) plus a manual-approval-gated deploy workflow for the optional AWS path, with the full dispatch → approval-pause → apply → destroy cycle verified end-to-end for real.
Next.js 15TypeScriptSQLiteDocker ComposeOpenTelemetry CollectorTerraformGitHub ActionsAWS CloudWatch/IAM
- Designed and built a contract-first OpenAPI testing platform from scratch as a TypeScript pnpm monorepo (9 packages/apps): wrapper API, contract runner CLI, drift detection engine, AI drift analyzer, MCP server, and mock server.
- Contract runner validates live API responses against an internal OpenAPI spec using AJV schema validation; distinguishes transport errors, upstream schema drift, contract violations, and test assertion failures — each reported separately.
- CI pipeline: type-check → unit tests → Docker container build → live contract validation; 24/24 unit tests and 8/8 contract tests passing with 0 violations confirmed in CI artifact.
TypeScriptFastifyZodAJVVitestDockerGitHub Actions
TestRail MCP — MCP Server for TestRail Test Management
- Built a from-scratch MCP server wrapping TestRail's REST API v2 (projects, suites, sections, cases, runs, plans, tests, results), mirroring the dual-transport architecture (stdio locally, streamable-HTTP on Google Cloud Run) and OAuth 2.0/PKCE + signed JWT security model of the UseResponse MCP.
- Two-tier tool-gating scheme: five category flags (read, case/run/result writes, deletes) stacked with an optional named-tool allowlist that can only narrow — never widen — what a category flag permits; deletes deliberately scoped to case/suite only.
- Always-on list_tool_policy introspection tool reports live category flags and resolved tool list; automated test suite validates registered tools match policy across representative flag combinations.
PythonFastMCPhttpxDockeruvicorn/Starlette
- Architected a 5-phase AI-driven test pipeline: Planner (PRD → structured test plan), Generator (Playwright TypeScript specs), Debug (resolves real failures — config mismatches, assertion brittleness), Healer (simulates selector drift and auto-repairs broken locators).
- 39 Playwright TypeScript tests across 5 suites, all passing; real bugs diagnosed and fixed during the debug phase.
- Demonstrates self-healing automated integration test infrastructure — locators repair themselves on UI changes, eliminating manual maintenance.
PlaywrightTypeScriptClaude Code
- Built a reproducible Kubernetes CKA exam preparation platform on a local kind cluster with deterministic task verification and a local AI agent solver requiring no cloud API key.
- Task system: each task injects a broken Kubernetes state (misconfigured Services, missing RBAC, stuck PVCs, scheduling taints); a deterministic verify.sh confirms the fix — no fuzzy grading.
- AI solver: local Claude Code agent explains each kubectl command and interprets output at every step; 5/5 tasks passed autonomously in a single suite run.
BashPythonPowerShellkindkubectlClaude Code
- One-button GitHub Actions CI/CD deploying production Kubernetes on OCI Always Free Tier — $0/month vs $250+/month equivalent on AWS.
- Continuous testing infrastructure: Prometheus + Grafana monitoring, automated certificate management, WireGuard VPN zero-trust access.
- Claude Code integration: 5 custom Kubernetes skills and MCP server for natural-language cluster management.
TerraformAnsibleKubernetes (OKE)HelmGitHub ActionsWireGuardPrometheusGrafana
- Developed a Next.js 15 single-page template for professional and business use (TypeScript, Tailwind CSS v4, shadcn/ui, Credly certification badge integration), with environment-based builds configured for Meta Business Verification readiness — the platform this multi-facet site itself is built on.
Next.js 15TypeScriptTailwind CSS v4shadcn/ui
Skills
Test Automation
PlaywrightAPI testingAI-augmented test generationself-healing locatorsPostman
Testing Methods
Automated integration testingend-to-end testingregression testingschema/contract validationexploratory testingcontinuous testingAgile/ScrumISTQBshift-left
AI & Agentic Systems
Multi-agent pipeline designClaude CodeMCP server developmentOAuth2/JWT-secured streamable-HTTPruntime capability discovery (Unreal Engine 5.8)vision-based UI agents
CI/CD & DevOps
GitHub ActionsDockerKubernetesHelmTerraform
Workflow Automation
n8nmulti-service orchestrationApifyschema validation
Monitoring
OpenTelemetryELK StackPrometheusGrafanamulti-platform log aggregation
Tools
TestRailJIRAGit/GitHubREST APIsNode.jsTypeScriptPostgreSQLSupabase
Experience
Automation & Integration Engineer — Freelance
2023 — Present- Built AI-driven agentic workflows including a RAG chatbot (OpenAI, Gemini, Supabase vector DB) connected to WhatsApp and Telegram via webhook integrations.
- Developed accounting workflow automation: web form → n8n webhook → data processing → government-regulated XML/PDF, with XML schema validation against Czech regulatory requirements.
- Engineered automated data extraction, processing, and media handling pipelines for multiple clients (details under NDA).
Advanced Quality Assurance Specialist | Junior Release Manager — Keen Software House a.s.
Aug 2020 — Jul 2026- Architected SE Tester AI Agent System — a three-agent pipeline (Planner, Generator, Healer) that produces and self-heals XML test scripts for the internal Space Engineers headless/headed test tool.
- Designed and maintained automated integration test workflows covering API validation, external service integrations, and Apify data pipeline schema validation.
- Operated a multi-platform log ingestion system (ELK stack) aggregating telemetry from Xbox, PlayStation, and Steam for centralised QA verification and anomaly detection.
- Embedded across the full QA lifecycle for 13 DLC releases (Wasteland through v210): design doc review, pre-release testing, Steam release cycle support, post-release patch triage.
QA Engineer — Charged Monkey
Oct 2017 — Aug 2020- Sole QA for most of the tenure, owning end-to-end test strategy, automation pipeline, and the full release cycle across iOS (App Store), Android (Google Play), and Amazon/Kindle Fire platforms; briefly led a small, two-person QA team.
- Established QA practices from the ground up; mentored incoming team members.
- Defined and executed test strategies covering automated regression, functional, integration, and platform-specific testing.
Certifications
Oracle Cloud Infrastructure 2025 Certified Generative AI ProfessionalOracle Cloud Infrastructure 2025 Certified Architect AssociateOracle Cloud Infrastructure 2025 Certified Foundations AssociateCertified in Cybersecurity (CC) — ISC2ISTQB® Foundation LevelAPI Test Automation