eugene-kim

Agentic coding loops work well when the LLM can run tests to verify its work. For backend code, this is straightforward - write a test, run it, iterate. For front-end work, I haven't found a good workflow. I've tried…