Building Scalable Test Frameworks for Robust Software Delivery.
Bridging technical excellence with human connection to build resilient teams and robust software.
I believe quality isn't just about finding bugs—it's about building confidence in every release. I currently run two concurrent senior engagements: sole architect of the QA stack at TOKERO (Playwright .NET, NBomber, custom Blazor reporting) and automation lead for Deutsche Bahn's SAP ERP Integrated Railway Management System at Heaven Solutions (Karate API, Playwright UI). I thrive where technical precision meets collaborative problem-solving.
My approach combines cutting-edge automation tools with practical team dynamics, and increasingly with AI-augmented workflows— Claude Code skills, MCP integrations (Playwright, Supabase), and spec-driven automation that scales across QA engineers. I'm passionate about transforming testing bottlenecks into quality accelerators and making quality everyone's responsibility.
I find bugs before users do
Faster releases, fewer bugs
I elevate everyone around me
Architecting frameworks adopted as team standards across multiple client engagements
Mentoring junior QA engineers on test design and automation best practices
Exploring testing innovations — currently building custom Claude Code skills, Playwright MCP integrations, visual regression workflows, and subagent-driven test development
• Want to discuss API testing strategies vs. UI automation trade-offs
• Need a partner to transform "testing bottlenecks" into "quality accelerators"
• Believe teams deliver better software when quality is everyone's responsibility
Open to collaborations, mentorship, and conversations that strengthen software quality practices.
Hours Worked
🔴 Live Counter
Projects
🚀 Delivered
Frameworks
⭐ Created
Courses
🎓 Completed
Automation isn't about replacing humans—it's about giving us time to be more human.— Personal Philosophy
Shift-left testing initiative integrating quality into development pipeline
Reduction in defects
Playwright E2E testing with parallel execution and auto-waiting
Faster testing cycles
Karate Framework data-driven testing for microservices validation
Less maintenance
Jira/XRay integration connecting tests to requirements
Quality dashboards
Sole architect and maintainer of the QA stack at TOKERO (European crypto exchange): a Playwright functional framework in production since 2025, plus an NBomber performance suite and a custom Blazor reporting platform — both shipped to production in 2026. Owned end-to-end since July 2025.
Prevented €2M+ in potential system failures through proactive defect detection for Europe's largest transportation network. Reduced testing cycles by 40% while maintaining zero-tolerance safety standards for 2+ billion annual passengers. Led digital transformation of critical railway infrastructure supporting Germany's €40B transportation modernization program.
Led precision testing for medical manufacturing systems ensuring ±0.001mm accuracy standards. Delivered 100% on-time deliverables as interim team lead during critical product launches. Resolved 65+ critical defects with 99% clarity, reducing developer resolution time by 40% for CAD/CAM workflows.
Increased client acquisition rates by 35% and reduced manual work by 80% through quality-focused development. Maintained 99.9% uptime for Romania's leading advertising agency. Built scalable solutions that transformed 100+ SMEs from local businesses into digital market leaders across Eastern Europe.
How I architect end-to-end QA stacks — design patterns, layering, and tooling choices. Generalized from production work (business logic, scenario names, internal endpoints, and customer-specific data omitted).
Playwright .NET / C#
Page Object Model with HybridFixture base; lazy-loaded POs; fixture lifecycle managed centrally.
Cross-browser, multi-environment (4 tiers), 10-locale support, screenshot / video / HAR capture, DB validation via EF Core.
Custom KPI tracking: TTFB, LCP, FCP, CLS. CSV / HTML exports for trend analysis.
GitLab CI with parallelized execution; Azure Key Vault for secrets; source generator for route constants.
qaatpw/ ├── src/ │ ├── Fixtures/ // HybridFixture + TestBase │ ├── Pages/ // 77 Page Objects (by domain) │ ├── Tests/ // 49 test classes │ ├── Helpers/ // Assert · Blazor · DB · Data │ ├── Workflows/ // End-to-end user journeys │ ├── Metrics/ // TTFB · LCP · FCP · CLS │ ├── Config/ // Per-env + Azure Key Vault │ ├── Data/ // Localization (10 locales) │ └── Prompts/ // AI workflow templates └── Tokero.Tests.slnx
NBomber 6.x · .NET 10
API endpoints, endurance soaks, SignalR / WebSocket handshake (independent-circuit), saturation curves.
Client pool management, custom NBomber sinks, auth semaphores, configurable error class taxonomy.
30+ ready-made profiles: Quick smoke, Progressive ramp, CDN-safe, Standard (origin), Stress / Spike.
Grafana dashboards, Prometheus metrics, Sentry error tracking, pod-distribution tracking for HPA validation.
perf/ ├── Scenarios/ │ ├── Api/ // API endpoint scenarios │ ├── Endurance/ // Soak · memory leak │ └── SignalR/ // WebSocket handshake ├── Infrastructure/ │ ├── ClientPools/ // HTTP client mgmt │ ├── Auth/ // Tokens + semaphores │ ├── Sinks/ // Custom NBomber sinks │ ├── Diagnostics/ // Pod + CPU/Mem tracking │ ├── Grafana/ // Dashboard provisioning │ └── Sentry/ // Error tracking ├── Profiles/ // 30+ load profiles ├── Dashboard/ // Separate Blazor app ├── TestData/ // Sanitized fixtures └── Tokero.PerfTests.slnx
Blazor Server · .NET 10
4-tier solution: Core (zero external deps) / Data (EF Core, repos) / Web (Blazor Server) / Tests.
MudBlazor components, Blazor-ApexCharts for candlestick + trend visualizations, three-zone Duration page.
Dual-provider data layer: SQLite for local dev, PostgreSQL (Supabase) for production — zero schema divergence.
NUnit XML parsing pipeline, flaky-test detection, duration regression at 10% threshold, configurable data retention.
pulse/ ├── TestReportingTool.Core/ // Domain (zero ext deps) │ ├── Models/ │ ├── Interfaces/ │ ├── Services/ │ ├── Constants/ │ └── Utils/ ├── TestReportingTool.Data/ // EF Core + repositories ├── TestReportingTool.Web/ // Blazor Server UI │ ├── Components/ // MudBlazor + ApexCharts │ ├── Controllers/ │ └── Services/ ├── TestReportingTool.Tests/ // xUnit + Moq └── TestReportingTool.slnx
Karate API regression suite (Java 17) — 264 feature files across 17 microservice domains (billing, train-path, master-data), exercising 23 services.
Playwright UI framework (TypeScript) — 22 page objects with separated locator classes, 152 specs / 254 tests, plus typed OpenAPI clients for 44 services.
CI integration — GitLab CI: Karate runs a 23-job sequential regression pipeline; reports to GitLab Pages and synced into Jira / XRay.
Test data & infra — multi-database setup (PostgreSQL / MySQL / Oracle) with Kubernetes secret injection; 134 SQL scripts plus Excel/JSON fixtures.
act-backend-tests/ // Karate API · Java 17 ├── src/test/java/ │ ├── karate-config.js // env switch (tu/abn/e2e/eu) + URLs │ ├── services/ │ │ ├── BaseTest.java // parallel run + Cucumber reports │ │ ├── Run*.java // 22 JUnit5 runners (by @tag) │ │ └── 17 domains/ // 264 .feature files │ └── logback-test.xml // logging config ├── templates/ // GitLab CI · 23 sequential jobs │ ├── base_pipeline.yml // stages · secrets · job templates │ └── test-env-*.yml // TU / ABN / E2E triggers ├── scripts/ // xray-post · xray-attachment · merge-xml └── pom.xml // Karate 1.4.1 · JUnit5 · multi-DB drivers playwright-tests-automation/ // Playwright UI · TypeScript ├── pages/ // 22 page objects (BasePage) ├── locators/ // 17 locator classes (selector-only) ├── clients/ // openapi-fetch client factory ├── generated/ // 12 OpenAPI → TS type files ├── config/ // 7 environments · 44 services ├── helpers/ // fixtures · DB · SQL · multi-user ├── sql/ // 18 reusable SQL scripts ├── data/ // 63 fixtures (xls/xlsx/json/pdf) ├── tests/ // 22 areas · 152 specs · 254 tests │ ├── auth.setup.ts // KeyCloak auth project │ └── …/ // specs by business area ├── templates/ // CI + secret management └── playwright.config.ts
Real two-repo split (Karate API + Playwright UI); structure condensed, internal names sanitized.
9+ custom Claude Code skills for standardized Page Object & test generation (e.g. /generate-page, /generate-test, /verify-test, /debug-failure).
Playwright MCP integration — agentic UI verification of new tests interactively before commit.
Subagent-driven roles — planner / generator / reviewer separation for higher-quality artifacts.
Plan-first / spec-driven — flow goes specs → executable plans → implementation, keeping tests aligned with intent.
qaatpw/.claude/ ├── skills/ │ ├── generate-page/ // Scaffold PO from Razor │ ├── generate-test/ // Test class from PO │ ├── verify-test/ // Playwright MCP check │ ├── debug-failure/ // RCA assistance │ ├── review-test/ // Code review pass │ ├── playwright-devops/ // CI/CD assistant │ ├── docs-maintainer/ // Docs sync │ ├── runner-host-ops/ // Runner / infra ops │ └── skill-creator/ // Meta — builds skills ├── hooks/ // Pre/post commit hooks ├── team-memory/ // Shared team context └── settings.json // MCP + tool config
9 production skills active in the test repo — Playwright MCP + Supabase MCP wired in.
Certified Tester Foundation Level - Software Testing Fundamentals
Since July 2021
PSM I — Scrum framework fundamentals and Agile delivery practices
Scrum.org certified
Advanced E2E testing with TypeScript/.NET and parallel execution
3+ years hands-on
CI/CD pipelines, version control, and DevOps practices
Daily workflow
Docker containerization and Kubernetes orchestration
Production systems
Karate Framework, Postman, and microservices validation
3+ years hands-on
Exploring AI testing tools and visual regression testing
Future-focused
Advancing toward Tech/QA Lead through strategic test leadership, team mentoring, and enterprise-scale automation frameworks
Machine learning in test automation and predictive quality analytics
Advanced security testing techniques and OWASP methodologies
Interested in discussing QA automation opportunities or have questions about testing strategies? I'd love to hear from you.
I typically respond to messages within 24 hours. For urgent matters, LinkedIn tends to be the fastest way to reach me.