Note: Due to compute limitations of the server this website is hosted on, generation will take a few minutes.
SAM-1
A 9M parameter causal language model trained on the TinyStories dataset to complete short children's stories.
Noise Detection VAE
The presence of noise in a signal can render that signal useless. I solve this problem by implementing a variational autoencoder to encode electrocardiogram signals. I then use a neural network to classify the encoded data, achieving a balanced accuracy of 76% across 3 noise classes.
This Portfolio Website
This website was built to showcase my full-stack development skills. It also showcases my systems engineering and MLOps abilities by building a custom Docker container for this website software, as well as connecting it to my proprietary SAM-1 model for showcasing generation abilities. All of which is hosted on a single server for cost-efficiency.