Project
What To Test is a personal MVP for QA interview practice
The project helps QA candidates practice open-ended interview prompts by writing what they would test, comparing their answer with structured checklists, and reviewing junior, middle, and senior QA thinking.
Core features
- Interview-style QA scenarios for answering "How would you test this?"
- Concept-based local scoring with covered and missed risk areas
- Junior, middle, and senior model answers
- Optional OpenAI-powered AI Review after local scoring
- Guided external practice labs with original tasks and checklists
- Local progress tracking and lab notes using browser localStorage
Tech stack
- Next.js
- TypeScript
- Tailwind CSS
- Vercel
- Vercel Analytics
- OpenAI API
- localStorage
What this project demonstrates
Product thinking around a focused QA interview preparation use case
Structured content modeling for realistic test design scenarios
Server-side API integration without exposing secrets to the browser
Client-side state management for progress, notes, and MVP usage limits
Clean responsive UI with practical empty, loading, and error states
Current limitations
- No authentication or user accounts
- No database or server-side progress history
- AI Review is optional and experimental
- The browser-level AI Review limit is MVP cost control, not abuse prevention
- Scoring is still heuristic and should be treated as training feedback, not a certification result
Roadmap
- More curated interview packs and API test design prompts
- Better scoring calibration and answer quality signals
- Deeper weak-spot analysis across learning paths
- Mentor or classroom mode for assignments
- Exportable practice reports for interview preparation
Live demo
Try the product as a user would: start with Interview Mode, complete a challenge, review the local score, and optionally request AI Review if the deployment has an OpenAI API key configured.
Start interview practice