Quality Assurance Architect

Architect (Quality Assurance, Test Automation & AI-Assisted QA)

I transform the software development lifecycle within organisations and establish solid quality assurance strategies. My focus is optimising quality while reducing the cost of recurring engineering tasks through automation — from classical test automation to AI-assisted test generation, self-maintaining suites and voice-driven QA tooling. Quality Assurance is one of the most critical components of any IT business; automation is what makes it scalable. I build the frameworks, the agents, and the culture that deliver both.

✦ BLEEDING EDGE ✦

For the past year my primary focus has been designing and building a fully autonomous, agentic Quality Assurance platform — codename Magnum Opus. An end-to-end lifecycle in which AI agents discover an application, author its test suite, expand coverage, hunt bugs, triage failures and report to stakeholders — all with human-grade discipline and auditable artefacts. Built on my Civitas Cerebrum framework and orchestrated by a single project-level configuration that turns any repository into an autonomous QA target. See Magnum Opus section below. Internal mechanics intentionally kept under wraps.

CORE EXPERTISE

• QA Strategy & Leadership — Test Maturity Models, SDLC design, QA team formation, performance frameworks, supplier governance.
• Test Automation Architecture — Web, mobile (iOS / Android), desktop, API, database and email, unified under single cross-platform frameworks.
• AI-Assisted Quality Engineering — Staged autonomous agentic QA pipeline (journey mapping → authoring → coverage expansion → bug discovery → failure diagnosis), adversarial / red-team AI testing, and model-agnostic LLM orchestration.
• Open-Source Library Design — Published multiple Maven Central and npm libraries used in production testing ecosystems.
• CI/CD & DevOps for QA — Azure DevOps, GitHub Actions, GitLab CI, Jenkins, custom parallelisation, Docker, Allure / Calliope / Slack reporting integrations.
• Languages & Stacks — Java, TypeScript / Node, Python, Swift. Spring Boot, Playwright, Selenium, Appium, Cucumber, Retrofit, React, Vue, Next.js.

WORK EXPERIENCE

Quality Assurance Architect & Managing Consultant @spriteCloud (01.01.2024 - present)
SpriteCloud is a consultancy providing software testing services & test results dashboards.

• Assessed clients’ quality assurance procedures to develop Test Maturity Models (TMM).
• Designed new Software Development Lifecycle (SDLC) models for various clients that enabled an effective quality assurance discipline.
• Determined the necessary amount of QA resource per work stream while managing the QA team within organisations to maximise the effectiveness and efficiency of QA practices.
• Developed performance measurement strategies to ensure continuity of performance and career growth of QA Engineers, helping the group of consultants in my team establish KPIs.
• Trained & supported QA Engineers to excel at software development & test automation.
• Designed and rolled out a new SDLC for MR MARVIS, aligning development, QA and release practices into a coherent delivery model.
• Lead innovation as the Guild Lead of Innovation.
  • Designed and delivered a multi-model AI QA agent combining vision models, LLMs and structured prompting for autonomous test scenario generation and test maintenance.
  • Architected an easy-to-deploy CV solution modernising consultancy CV practices (the system rendering this document).
  • Delivered desktop-to-database end-to-end test automation to secure a major client engagement with Ahold Delhaize.
  • Led the design of AI-native test authoring workflows adopted internally by consultants, bringing LLM-assisted test generation into day-to-day delivery.
    
    

Quality Assurance Solution Owner @citizenM hotels (01.03.2022 - 31.01.2026)
citizenM is a Netherlands-based hotel developer, investor, and hotel chain.

• Tasked with inspecting the entirety of citizenM’s software development practices to develop a Test Maturity Model.
• Created an extensive TMM covering five stages of quality assurance maturity & uncovering potential optimisations.
• Took the initiative to develop a centralised QA approach & managed the QA transition within citizenM.
• Helped citizenM increase product quality by establishing quality standards while detecting points of improvement & vulnerabilities.
• Founded the QA team to increase quality standards across citizenM’s product lineup:
  • Managed the team providing QA to several product teams.
  • Conducted numerous hiring interviews & assessments to find the right talent.
  • Continuously tracked performance of team members.
  • Did resource reallocation to compensate for shifting test workload across different product teams.
  • Created an agile work environment & a continuous performance feedback cycle to ensure career development.
  • Trained team members on test automation.
• Redesigned the software development lifecycle to enable a solid test strategy & guided suppliers to migrate to the proposed SDLC model.
  • Created a functionality analysis on the existing SDLC model.
  • Designed a brand-new SDLC model compensating for its shortcomings.
  • Guided citizenM & its suppliers through the SDLC migration.
  • Organised the QA team around the new model.
  • Enabled a test architecture separating development, test and acceptance environments while integrating test automation in deployment pipelines.
• Led Quality Assurance through citizenM’s backend migration.
  • Handled strategy & budget planning.
  • Hired necessary resources.
  • Designed & implemented a performance test suite (JMeter, Octaperf).
  • Designed a data migration audit.
  • Defined a UAT strategy between citizenM & its suppliers.
• Created clean, efficient, scalable & unified test automation solutions for various teams.
  • API test automation for the entirety of citizenM’s backend services.
  • Web UI test automation for citizenM’s website.
  • Enabled cross-layer verification between frontends, backends, databases, customer emails and the analytics layer.
  • Established continuous integration using Azure DevOps.
  • Developed a custom parallelisation script enabling test parallelisation at the DevOps level.
  • Slack integrations: live pipeline monitoring, a results-posting bot with slash-command remote triggers, and Calliope.pro → Slack reporting.
  • Established a single test automation framework that simultaneously tested frontends, middleware, emails, analytics layer and the database against each other, enabling true E2E test automation — scaled with the QA team I founded to cover 5000+ test steps.
    
    

Senior Test Automation Engineer @spriteCloud (01.12.2021 - 01.01.2024)
SpriteCloud is a consultancy providing software testing services & test results dashboards.

• Expanded expertise in additional CI tools and automation frameworks including:
  • GitLab CI
  • Parallelisation with Knapsack (for Cypress, Jest, RSpec, etc.)
  • Automated vulnerability and security testing with ZAProxy & Burp Suite
  • Web UI automation with Cypress
  • Test automation with Jest
  • Automated bug reporting with Jira & Trello integration
• Published a Mobile UI Test Automation Template utilising my mobile automation library Mobileib, built on Java & Appium, capable of parallel test execution (multithreaded and/or via BrowserStack).
• Published an API Automation Template utilising my API automation library Api Assured, built on Java, Retrofit2 & OkHttp3, capable of parallel test execution, request interception & advanced header management.
• Improved upon the Web UI Automation Template & the web UI automation library Pickleib.
• Improved test result monitoring using Calliope.pro & Allure for more insightful reporting.
  
  

Software QA Automation Engineer @Getir (02.06.2021 - 01.12.2021)
Getir is a technology company that provides solutions for the delivery of goods in urban areas.

• Work involved web automation, BE automation & BE development with JPA and Spring Boot. Provided testing infrastructures for developers and testers, created virtual courier services with Docker, prepared easily triggered regression suites and wrote BE services for other teams.
• Single-handedly programmed the entire web UI regression suite for one of Getir’s sub-brands, with a scaling adaptation of the Abstract Factory pattern enabling the project to support multiple sub-brand domains.
• Set up CI jobs and used Portainer slaves to pick up available jobs from Jenkins.
• Prepared Quickstart libraries combining BDD (Cucumber & Gauge) and Page Object Model, allowing entire automation frameworks to be bootstrapped within minutes.
• Implemented parallel execution as standard on every project.
• Authored an experimental automation framework that automatically acquires all elements & attributes with auto-generated xPaths into a project JSON file — interpreted by the framework — allowing automation without any manual page inspection (the conceptual ancestor of my current JSON-repository-based frameworks).
• Helped over 50 QA engineers at Getir learn test automation through a three-week bootcamp. Sessions were captured and are reused for ongoing onboarding.
  
  

Software Test Automation Engineer @Pixery (15.07.2020 - 02.06.2021)
Pixery develops next-generation mobile applications for content creators. Worked on both Funimate (a social video editing platform used by millions of active users daily) & Impresso (an all-in-one video editing tool for influencers and businesses).

• Ensured end-product quality through comprehensive tests with creative test cases for each functional task, and integration tests before each release.
• Provided a flexible automation framework simultaneously usable across several mobile & web projects and API tests, with BDD scenarios writable by non-technical team members.
• Developed Swift-based mobile test tools used to measure the accuracy of the video editing engines.
• Deep experience with Java, Maven, TestNG, Cucumber, Selenium & Appium, and Jenkins CI (/w Groovy).
  
  

Software QA Engineer @JustSnap (04.10.2018 – 01.07.2020)
Justsnap provides campaign solutions based on its in-house receipt-scanning ML technology.

• Participated in development, QA and BA processes of campaign projects for P&G, Phillips and Pepsi.
• Responsible for quality assurance activities across 17+ campaign projects.
• Worked on cross-platform mobile applications; explored mobile automation with Selenium & Katalon Studio.
• Created test scenarios, performed sanity and regression tests.
• Implemented flexible web UI test automation solutions that were easy to set up, keeping pace with fast delivery schedules.
  
  

✦ MAGNUM OPUS — AUTONOMOUS AGENTIC QA PIPELINE ✦

The headline project of my current work. An end-to-end agentic Quality Assurance lifecycle in which AI agents discover an application, author its test suite, expand coverage, hunt bugs, triage failures and report to stakeholders — with human-grade discipline and auditable artefacts at every step.

The pipeline is staged, opinionated and verification-driven. A single project-level orchestration layer — codename Magnum Opus — sits atop it all and activates the full lifecycle for any repository it is dropped into. Internal mechanics intentionally kept under wraps.

Lifecycle at a glance

• Journey Mapping — Autonomous app discovery. The agent navigates the live application, identifies pages and user flows, and prioritises journeys by business impact.
• Authoring — Initial test authoring against the Singularity Steps API, with selector-free page objects sourced from a central JSON repository.
• Test Composer — Iterative coverage expansion. Cycles of scenario generation → implementation → stabilisation → review → gap analysis until the suite comprehensively covers the application.
• Bug Discovery — Adversarial quality audit. The agent probes the live app with fresh eyes (“first-time effect”), cross-references findings against accumulated context and the green test baseline, and produces a prioritised bug report with reproduction tests.
• Failure Diagnosis — Evidence-based triage for failing tests: screenshot + DOM analysis, root-cause hypothesis, autonomous fix or app-bug report with evidence.
• Maintenance / Onboarding / Assistance — Keeping suites healthy, bringing new codebases into the pipeline and everyday assistive tasks.

Specialist testing modules invoked per scenario: functional-testing, negative-validation, security, permission-access, responsive, state-session, search-filtering, usability-ux, compliance-review.

Adversarial & governance

• agents-vs-agents — Red-team / blue-team AI testing: guardrail verification, prompt-injection, bias detection and compliance auditing for any LLM-integrated product.
• work-summary-deck — Generates branded stakeholder decks summarising coverage, findings and progress.

Agent-engineering discipline

• small-model-agent — A strict PLAN → FOCUS → ACT → VERIFY loop that keeps ≤30B-parameter local models (via Ollama / llama.cpp) productive on real engineering tasks.
• agent-debug-mode — Structured logging of agent reasoning at token boundaries, enabling post-hoc analysis of agent behaviour.

Together these make autonomous QA reproducible, auditable and model-agnostic — the same pipeline runs under Claude, local Llama / Qwen models, or mixed setups, with consistent artefacts and reporting.

CIVITAS CEREBRUM — AI-NATIVE QA FRAMEWORK

The engine beneath the Magnum Opus pipeline. A modern ecosystem for AI-native test automation — published on npm under @civitas-cerebrum/*. The libraries separate selectors from test logic, unify execution across web / mobile / desktop platforms, enforce API test coverage, and give LLM coding assistants the structured primitives they need to author reliable tests.

element-repository

Platform-agnostic element registry. A single JSON file is the source of truth for all locators across Playwright (web) and Appium / WebDriverIO (mobile).

element-interactions

High-level semantic interaction & verification layer on top of element-repository, with a built-in Claude Code skill that teaches AI assistants how to author tests in the framework.

singularity-engine

Cross-platform test execution engine applying an Abstract Factory pattern to expose a single Steps API that runs unchanged on web (Playwright), Android, iOS, Windows and macOS (Appium). Includes first-class API and email test primitives. Distributed as an obfuscated public npm package.

test-coverage

Static-analysis tool (TypeScript AST) that enforces 100% API coverage in test suites, with CI-friendly reporters — GitHub table, HTML, JSON and badges.

email-client

Zero-dependency TypeScript SMTP/IMAP client purpose-built for E2E flows (signup → verification → click-through).

Jarvis

Local, privacy-preserving voice-driven QA copilot. Audio pipeline: Silero VAD → GTCRN denoise → WeSpeaker speaker verification → Whisper STT → LLM → Kokoro TTS. Exposes an MCP server so Claude Code can be driven entirely by voice.

PERSONAL OPEN-SOURCE LIBRARIES

Programming is my passion. I like to develop reusable modules of software which I then publish as open source for others to benefit from. Selected libraries below.

Pickleib

A utility library for software automation projects. It helps design and run tests with Selenium WebDriver in a simple and efficient way, providing driver setup, clean API capabilities, POM support, email test capability, database connection, a data layer and event tests.

Check out the latest version here.

Wasapi

A lightweight, Java-based API utility library that simplifies HTTP service generation and API calls using Retrofit. It abstracts the boilerplate needed to set up and use APIs, making it easier to integrate network calls in Java applications.

Features

• Simplified API service generation
• Centralized HTTP configuration
• Easy-to-use wasapi.Caller utility for handling requests
• Designed with modularity and readability in mind

Check out the latest version here.

Java Utilities

A collection of utility methods developed over the years. It includes object comparison methods, file manipulation classes for files, JSONs, Excel files and zipping, a custom basic logger, utilities for list, string and numeric manipulation, a built-in text parser, localisation, command execution capabilities & more.

Check out the latest version here.

Context Store

Utility classes for thread-safe context storage and secret property handling. Dual-platform — published on both Maven Central and npm.

Check it out here.

Ollama API

Integrate LLMs into your Java projects in seconds. Built to streamline LLM integration in Java — no manual JSON parsing or brittle API calls. Provide a prompt and a target class; the library returns structured, type-safe responses. Ideal for chatbots, AI tools, agents and internal LLM utilities.

Check it out here.

Easy Secret Santa 🎅🎁

A Java application designed to simplify and automate the process of organising a Secret Santa gift exchange. It pairs participants, generates personalised emails with recipient details, and sends them via SMTP — ensuring a smooth, festive experience. Originally built for my team, now used by multiple organisations.

Check it out here.

AUTONOMOUS AGENTS & AI DATA PIPELINES

Research projects that combine my own libraries (Pickleib, Wasapi, Ollama-API) with vision and language models to build end-to-end autonomous agents.

• Napoleon — Java-based autonomous web agent. Selenium-driven browser control through Pickleib, vision-based UI understanding via a custom Java client, and decision-making through my Ollama-API library. Features a dedicated Nexus reasoning core with structured action models, a Commandments policy layer and a ValidationResponse self-check loop — perceive → classify → decide → validate → act, fully in Java.
• T-800 — Autonomous agent focused on adversarial web navigation, including a vision-LLM-driven CAPTCHA resolver that crops candidate elements, queries a vision LLM for target indexes and assembles a verdict. Built on Selenium 4.19 and my Utilities library.
• EMTP (Expert Model Training Pipeline) — Python pipeline for producing training datasets for domain-expert AI models. Modular acquisition → enrichment → questions architecture, DuckDuckGo search with dorking support, Firecrawl-based markdown scraping, Unicode normalisation and Q&A-pair generation. Supports both interactive and non-interactive CLI execution.

PRODUCT & FULL-STACK PROJECTS

• BookHive — A production-grade full-stack bookstore (React 18, Spring Boot 3, MongoDB, Docker Compose) intentionally designed as an automation-friendly reference app: stable data-testid attributes, deterministic /api/seed and /api/reset endpoints, JWT + HttpOnly-cookie dual auth, and a real cart/order flow used as a canonical target for the Civitas Cerebrum ecosystem.
• Vue / Android / iOS Test Apps — A family of multi-platform reference applications (Vue 3 playground, native Android, native iOS) used as public testing sandboxes for validating cross-platform automation frameworks.
• Food Planner — Lightweight personal web app exploring productivity and lifestyle workflows.