Execution-Verified Reinforcement Learning for Optimization Modeling

Create an open-source framework for execution-verified reinforcement learning in optimization tasks. This allows developers to fine-tune smaller local models on verifiable solver outputs rather than synthetic LLM text.

Suggested repo: EVOM-lib

"Train smaller models to solve optimization tasks with verifiable execution feedback."

Estimated effort: 40h