Built a Hybrid NAS tool for RNN architectures (HyNAS-R) – Looking for feedback for my final year evaluation [R]

Hi everyone, I'm currently in the evaluation phase of my Final Year Project and am looking for feedback on the system I've built. It's called HyNAS-R, a Neural Architecture Search tool designed to automatically find the best RNN architectures for NLP tasks by combining a zero-cost proxy with metaheuristic optimization. I have recorded a video explaining the core algorithm and the technology stack behind the system, specifically how it uses an Improved Grey Wolf Optimizer and a Hidden Covariance proxy to search through thousands of architectures without expensive.