QA Automation Engineer based in Melbourne with close to three years of experience building reliable automation, improving software quality, and exploring practical applications of AI in testing.
About
I’m Sameer, a QA Automation Engineer based in Melbourne. I focus on building reliable software by reading tickets, tracing flows, and understanding how a system actually behaves before I automate anything against it.
I’m drawn to the practical use of AI in testing and spend time exploring where it can genuinely improve quality engineering. I care about solving problems well and continuing to learn as the craft evolves.
Experience
Software Test Engineer
Whistleblowing & compliance SaaS platform. Responsible for end-to-end, regression and exploratory testing, API validation and automation across environments.
Simplified and redesigned a critical regression suite by removing redundant cases and improving maintainability, then automated the cross-system workflows and cut a full day of manual testing by 2–3 engineers down to ~15–20 minutes.
Built and maintained Cypress UI and API automation across interconnected product systems and workflows, validating releases end to end rather than in isolation.
Contributed AI-assisted workflows for test design, analysis and documentation, alongside traditional automation.
Projects
Three projects that show how I think about testing, tooling and applying AI to real quality problems. Tap a card to jump to its case study, or keep scrolling.
A modular AI-powered QA engineering agent that orchestrates software testing workflows through structured, explainable, and skill-based execution.
Problem
QA investigations are fragmented across multiple tools and manual workflows. Engineers spend significant time gathering context, analysing failures and documenting findings. This work is repetitive, error-prone, and hard to hand off between people.
Solution
A modular agent architecture where each skill has a single responsibility, with explicit hand-offs and approval points between them. Instead of a single opaque model call, TestPilotAI chains focused skills that each produce a reviewable artifact, so the workflow stays explainable end to end.
Flow
Architecture & design decisions
Impact
A working proof of concept validating that modular AI workflows are a viable pattern for QA engineering, especially when the goal is trust and repeatability rather than raw automation.
Reflection
The challenge was designing trustworthy engineering workflows rather than simply integrating AI.
Redesigned and automated a critical regression workflow, replacing roughly a full day of manual testing by 2–3 engineers with ~15–20 minutes of automated execution.
Problem
A critical regression area required 2–3 engineers spending roughly a full day manually validating workflows across interconnected systems before every release. The TestRail suite itself had also grown redundant and inconsistent, which made it slow to run, harder to trust, and expensive to maintain.
Solution
Before writing any automation, I simplified the regression suite itself by removing redundant cases and tightening preconditions, steps and expected results so the suite reflected how the product actually behaved. Only then did I automate the refined workflows with Cypress, executed through AWS CodeBuild and Cypress Cloud.
Flow
Before · ~1 day · 2–3 engineers
After · ~15–20 min · automated
Architecture & design decisions
Impact
Reflection
Good automation starts with a good testing process, not just good code.
Exploring whether AI can generate higher-quality test cases by understanding broader system context, not just user stories.
Problem
Most AI-generated test cases work from a single user story. They often miss regression impact, edge cases, and system-level interactions that only become visible when you consider the wider context around a change.
Solution
The agent receives Jira tickets, user stories, linked design artefacts, existing TestRail cases and system knowledge files. From that context it proposes new scenarios, happy paths, negative and edge cases, and flags existing test cases likely to be impacted.
Manual review remains mandatory before finalisation.
Flow
Context inputs
Agent + review
Architecture & design decisions
Impact
Current status: ideation. Architecture and workflow are being engineered; there is no shipped product yet. The intent is to validate the core question before building further.
Reflection
Can richer system context help AI identify regression impacts that traditional AI-generated test cases often miss?
Contact
Open to QA Automation and quality engineering roles in Melbourne and remote. The fastest way to reach me is email.