Trained a custom 1D-CNN transformer architecture in PyTorch with spectral wave decomposition, paired with a Next.js visualization interface for researchers.
Epilepsy researchers face a major bottleneck in manual signal inspection. A graduating biomedical engineering capstone team partnered with Hopfield Labs to build an automated neural classifier.
We guided the research formulation, implemented Morlet wavelet transforms for time-frequency spectrogram extraction, and built a custom 1D convolutional vision transformer.
The student team received a Grade A+ defense evaluation, an IEEE conference publication acceptance, and open-sourced their benchmark weights.