Developed a toolkit combining surrogate neural networks with Particle Swarm Optimization to accelerate aerodynamic design cycles by 3x. The system provides visual insight into solver behavior, making optimization decisions transparent and tunable.
Result: Reduced design time by 66% and improved optimization efficiency by 10%.