This web-based lab lets you build, train, and explore Convolutional Autoencoders (CAEs) right in your browser. Configure your own architecture (dataset, bottleneck size, filters), watch epoch-by-epoch training curves, and dive into the encoder and decoder internals. Draw or upload images, inspect activation maps, tweak the latent bottleneck with sliders, and see real-time reconstructions. Designed to deepen your intuition about how autoencoders learn compressed representations, it’s a hands-on tool inspired by Dr. Sharma’s teaching philosophy.