Retrieve, Refine, Verify: Inverse Airfoil Design for Small Drones Without CFD

Problem: Designing low-drag wing shapes for drones and EVs normally needs CFD simulations that can take hours or days for every design change. That makes it too slow and expensive to adapt shapes quickly to new flight conditions. Solution: I built a free, pure-NumPy "retrieve, refine, verify" pipeline. It starts from similar airfoils already in a dataset and polishes them with a fast learned model until they hit a requested lift and drag. Every design is checked by an airfoil solver, and the pipeline returns designs in seconds. Between the training grid points it reaches 73% of targets in an XFOIL check, against 25% for a plain lookup.