AI Coding Benchmark
Developed production-like frontend engineering benchmarks used to evaluate large language models on realistic web development tasks and software engineering workflows.
Project Detail
This project does not have local screenshots in the project photo folder yet, so the detail page focuses on the case-study notes below.
Case Study Notes
Created reproducible development environments, task specifications, and reference implementations.
Automated functional and visual verification using Playwright, Puppeteer, and browser-based testing.
Improved benchmark reliability by eliminating ambiguous prompts and strengthening evaluation criteria.