ReVEL: Multi-Turn Reflective LLM-Guided Heuristic Evolution via Structured Performance Feedback

Build an iterative, multi-turn optimization framework that uses performance feedback to refine heuristics code. This moves beyond one-shot prompting to real-time heuristic improvement loops.

Suggested repo: revel-opt

"Let your LLM evolve its own algorithms through iterative performance feedback."

Estimated effort: 120h