Software Testing Consulting
I help teams build solid QA processes, test automation frameworks, and CI/CD pipelines — from zero to production-grade, the way I did at FedEx, Shell, Miro and Vodafone.
Hatem Hatamleh
Software Testing Expert & Agentic AI Engineer
16+ years building QA processes, test automation, and CI/CD pipelines for FedEx, Shell, Miro and Vodafone — now building and evaluating LLMs, RAG systems and Agentic AI. Founder of QAcart, 20,000+ students trained.
Vandebron Services
I provide a range of services across software testing and AI engineering — and I'm always excited for the next project. Whatever stage you're at, there's a way I can help.
GenAI & Agentic AI Evaluation
Expert evaluation of LLMs, RAG systems, Agentic AI applications and MCP servers — using DeepEval and custom test datasets to catch what manual review misses.
Training & Workshops
Online and on-site training for individuals and teams on software testing, test automation, Agentic AI and Vibe Coding — tailored to your level. 20,000+ students trained through QAcart.
AI Engineering
I build LLM applications, RAG systems, Agentic AI apps and MCP servers with LangChain, LangGraph, LangSmith, FastAPI and Python — from prototype to production.
Not sure which one fits, or have a question?
Contact MeNumbers I'm proud of — and I'm just getting started.
16+ Years in software testing
20K+ Students trained
11+ Freelancing assignments
7+ AI agents built
55+ Training workshops
11K+ YouTube subscribers
11K+ LinkedIn followers
22+ Speaking engagements
About Me Software Testing Expert.
Software Testing Expert.
GenAI & Agentic AI Engineer.
Senior SDET with 16+ years building QA processes, test automation frameworks, and CI/CD pipelines for teams at FedEx, Shell, Miro, Vodafone, and more. Today I bring that same testing rigor to Generative AI — evaluating LLMs, RAG systems, Agentic AI applications, and MCP servers with DeepEval and custom test datasets.
I'm also a hands-on AI Engineer: I build LLMs, RAG pipelines, Agentic AI applications, and MCP servers with LangChain, LangGraph, LangSmith, FastAPI, and Python — like the multi-agent platform powering QAcart. Jordanian and Dutch, splitting my time between the Netherlands and Jordan, and founder of QAcart with 20,000+ students trained in testing and Agentic AI.
- Based between
- Netherlands & Jordan
- Nationality
- Jordanian · Dutch
- Languages
- English Arabic Dutch
Case Studies
Five systems I built and shipped that teams and customers still use every day.
Business Rules RAG System
Shipping business logic is brutally complex — every country has its own rules, required documents and eligible services, and all of it shifts with the package, its contents and the service type. No single person could hold it. I built a RAG system over the codebase, business PDFs, years of Jira history and every country rulebook using LangChain and LangSmith, then connected Dynatrace, Postgres, Figma and Optimizely so it reasons about the live system rather than documents alone. It debugs production issues in seconds across the knowledge base, database and traces; refines any story and flags the affected areas and risks; and answers in moments the questions POs used to wait days for.
Happy Robot
An AI agent embedded directly in the shipment form. Customers chat with it or call it by voice and get answers grounded in real shipping rules — and it reads what they have already filled in, corrects errors, flags missing documents or information, and generates the paperwork they need, invoices included. I owned the evaluation: manual testing, DeepEval suites and prompt-injection probing to prove the agent answers correctly and never hallucinates. It now serves every customer shipping from the US to Canada, with more regions rolling out.
Recharge Mobile App
A React Native app serving 14+ European countries: every charge point on a live map with navigation, search and filtering, charge-session management and vehicle registration. Automating it was brutal — charge point availability, connector statuses and live sessions all shift under you, so nothing was reproducible. I moved the whole strategy left: a mock server built on NG API Mock reproducing every scenario — available, charging, faulted and the rest — which let us test the front end in complete isolation with Maestro, plus Testcontainers to cover the NestJS microservices on the backend. A handful of real transactions stayed on production as the final signal. The result is a large, stable test gate that runs in the pipeline before every release.
…and many, many more I'd love to walk you through.
Contact Me
Whether it's strengthening quality in an existing product, building test automation from the ground up, evaluating an AI system, or training your team — I'd be glad to hear what you're working on. Reach out through whichever channel suits you best.
Email hatem.hatamleh@gmail.com Write to me WhatsApp +31 6 27 004 821 Message me LinkedIn /in/hatem-hatamleh Connect with me GitHub /Hatamleh See my code
I reply to every enquiry within one business day.
Available for work in
Netherlands Belgium Germany Jordan UAE Saudi Arabia United States
Open for projects, assignments and trainings — remote worldwide, on-site across these countries.
FAQ
Two things that reinforce each other. I build and fix quality — QA processes, test automation frameworks, CI/CD pipelines. And I build and evaluate AI — LLM applications, RAG systems, Agentic AI apps and MCP servers. Most clients hire me for one and end up using both.