A team ships weekly, but each release needs two days of manual testing and bugs still reach customers. You build an automated suite that covers the checkout flow and API, runs on every pull request, and tells the team clearly when a release is safe.
A QA or SDET engineer makes sure software works and keeps working. You design test strategy, write automated tests as code, and find the risks others miss. In AI products you also test output that is not the same every time.
Goal: Learn how software is tested: write test cases, find and report bugs clearly, and write your first automated tests in a language you learn from scratch.
Before you start
Beginner · 8 weeks
Week 1: What testing is
Quality versus testing · Verification versus validation · Bugs, defects, failures · Types of testing · Vocabulary: test case, regression, severity
Practice: Test a website you use and write five clear bug reports.
Also: Vocabulary: test case, regression, severity
Practice: Test a website you use and write five clear bug reports.
Common pitfalls
Think first, then reveal
What is boundary value analysis?
Think first, then reveal
What makes a good bug report?
Think first, then reveal
Why are fixed sleeps bad in automated tests?
Think first, then reveal
How do you test something with non-deterministic output?
List the risks, decide what is tested at unit, API and UI level, and agree what must pass before release.
The plan for what to test at which level: many fast unit tests, fewer API tests and a small number of UI tests.
Why this pick
Putting every check in slow UI tests makes the suite flaky and too slow to run on every change.
What the role must know
Beginner mistake
Core
Automation
AI testing
Think first, then reveal
What is boundary value analysis?
Think first, then reveal
What makes a good bug report?
Think first, then reveal
Why are fixed sleeps bad in automated tests?
Think first, then reveal
How do you test something with non-deterministic output?
Think first, then reveal
How do you decide what to automate?
Think first, then reveal
Your suite is flaky. What do you do?
Think first, then reveal
What is a contract test?
Think first, then reveal
How do you evaluate a model-graded check?
Think first, then reveal
How do you decide how much testing is enough?
Think first, then reveal
How do you make AI-generated tests trustworthy?
Think first, then reveal
What metrics would you avoid?
Think first, then reveal
How do you keep an eval set from going stale?
Build one, measure it and put it on your CV. Each has tools, setup commands and milestones.
CV, portfolio and interview loop checklist →