← Projects

AI-assisted test infrastructure

S.M.A.R.T.

A team project with CHECK24 combining natural-language Playwright generation, deterministic mock data and execution feedback.

Student team - frontend, workflow & validation
  • TypeScript
  • Go
  • S3
  • Parquet
  • Redis
  • Docker
Product loop

Not just prompt in, test out

The interesting part is the feedback loop: a user describes test intent, generated Playwright code is validated against deterministic mock data, execution returns feedback, and the flow can be refined.

My contribution

Interface, validation and data handling

My work crossed UI flows, prompt validation, storage-connected paths and documentation. That makes the project useful in the portfolio as AI-assisted developer tooling, not as another generic chatbot example.

Execution feedback loop

From test intent to reproducible run

S.M.A.R.T. is best explained as a loop, because the value is in connecting generation, mock data and execution feedback.

01Natural-language intent
02Prompt validation
03Playwright generation
04Mock data
05Execution
06Feedback
07Refinement
08Documentation

Roadmap

Project lens

Built

Intent

User-facing flow for describing test goals and validating generated output.

Built

Data

Mock-data paths and storage-connected handling for reproducible test contexts.

Built

Execution

Playwright-centered feedback loop around generated tests.

Portfolio

Showcase

Demonstrates AI-assisted developer tooling with real workflow constraints.

Portfolio angle

What to show next

  • A short demo should show prompt, generated Playwright test and execution result in one loop.
  • The case page should stay focused on workflow, not on listing every backend dependency.
  • Screenshots from the UI will matter more than extra architecture text.