automated detection of defects in a running freight train

this was my master’s thesis at IISc where i trained different CNN models to detect defects in different parts of the train and they worked pretty well in the production

apparel recommendation system

developed a system that recommends fashion apparel based on user feedback

GANs from scratch

here’s my attempt at implementing DCGAN, LSGAN and BiGAN from scratch using tensorflow to generate new bitmojis

VAEs from scratch

another one from the “from scratch” series where i implemented vanilla vae and beta-vae using PyTorch

it's possible that there are more and i'm too lazy to put them here