Muhammad A.

Agentic QA and AI automation engineer, working as a forward-deployed (FDE) QA engineer: I find the bug, trace the root cause, ship the fix and add the test that keeps it fixed. I test LLM apps and AI agents, run QA with AI agents under my review, and build n8n and AI automations that are tested like software. 7+ years across fintech, core banking, digital health and e-commerce, including 8,000+ automated test cases on a live bill-payment platform. 🎯 WHERE CLIENTS USUALLY BRING ME IN 1. You built an AI chatbot, agent or LLM feature and need to know its answers are right before customers rely on it 2. Releases go out without a regression safety net, or the test suite is flaky and nobody trusts it 3. Bugs keep coming back because nobody owns the fix and the test together 4. Your team still copies leads, data or updates between tools by hand 5. An automation or integration fails silently and leads get lost 6. You want an AI assistant or agent to work with your own systems: your CRM, database, internal APIs or docs 🧠 MY AGENTIC QA WORKFLOW (I CAN SET IT UP IN YOUR PROJECT) AI agents do the repetitive QA work and I review everything that matters: 1. An agent reads your requirements, tickets or user stories and drafts the test cases 2. It writes and updates Playwright tests, and repairs them when the UI changes 3. The tests run in your CI on every pull request 4. Each failure is triaged as a real bug, a flaky test or an environment issue 5. Real bugs become tickets with steps, screenshots and logs, or I fix them directly I review every test and every bug before it reaches your team. 🛠️ WHAT I DO 🧱 Tests across the whole pyramid Unit and integration tests at the base, API tests in the middle and a lean set of end-to-end tests on top, so suites stay fast and failures point to the exact layer that broke. 🤖 LLM and AI testing Evaluation suites with DeepEval, golden datasets, hallucination and accuracy checks, prompt-injection testing and regression runs across model updates. I've worked with Vertex AI and Labelbox on model evaluation and training data quality. 🧪 Test automation Playwright and Cypress for web, Selenium and SpecFlow BDD for enterprise, Postman and REST Assured for APIs, wired into GitHub Actions or Azure DevOps. My pipeline-integrated suites cut deployment errors by 30%, and I raised API coverage by 40% on a multi-product SaaS platform. 🔍 Manual and exploratory testing Test plans, regression cycles, UAT, and bug reports a developer can reproduce on the first read. Real devices: Samsung Galaxy A07, Galaxy Tab A9+, iPhone 14 Pro and iPhone 16 Pro, plus BrowserStack and LambdaTest. ⚡ Performance testing Load and stress testing with JMeter and k6, with findings written for engineers. ⚙️ n8n and AI automation Lead capture into your CRM, AI agents that research, draft and classify, AI voice agents, and approval steps in Slack. Every workflow has dedupe, retries, error logging and alerts. 🔌 AI integrations I connect LLMs (OpenAI, Gemini, Claude) to the tools you already use: HubSpot and other CRMs, Slack, Gmail, calendars, Twilio, databases and your own APIs. Structured outputs, guardrails, retries and cost limits are built in, and every integration ships with tests. 🧩 Custom MCP server development I build Model Context Protocol (MCP) servers so Claude, ChatGPT and other AI agents can use your internal systems safely: scoped tools, authentication, input validation, audit logs and a test for every tool. 🚀 RECENT AI BUILDS (LIVE DEMOS IN MY PORTFOLIO) 1. Speed-to-lead SMS agent (n8n, Twilio, calendar booking, HubSpot, Slack). It texts back every missed call in under a second, qualifies the job, books it and passes 150/150 DeepEval checks. 2. AI recruiter outreach agent (n8n, Gemini, Gmail, Slack, HubSpot). It writes personalised emails, gets approval in Slack, classifies replies and stops follow-ups when someone answers. 3. AI landing page workflow (Claude Code). It builds a page from a reference and brand rules, checked by 23 automated QA tests. 4. AI lead engine (n8n, Postgres, Apollo, Hunter, HubSpot, calendar booking), in pilot. It researches companies, writes outreach, books meetings and keeps the CRM in sync. 🏢 WHERE I'VE DONE THIS TPS Worldwide (fintech bill payments), Meezan Bank (led Visa Direct integration testing with the Visa and Avanza teams) and Folio3 Software. I've trained internal QA teams and documented 15 critical bugs in a single audit cycle. 🏆 CERTIFICATIONS 1. ICT Quality Assurance Engineer, Australian Computer Society (ACS) 2. ICT Support and Test Engineers, ACS 3. Software Tester, ACS 4. PSEB Registered Freelancer, Pakistan Software Export Board ✅ Available now, 30+ hours a week. Send me what you're building and what's breaking, and I'll tell you honestly whether I'm the right fit.