A CNN+AE TensorFlow pipeline to recreate a patient's vision from an fMRI dataset.
A beat tracking implementation using PyTorch and TCNs.
An audio search re-implementation of the original Shazam audio identification system.
An investigation on combining a source separation model (spleeter) and a chord recognition model (a custom CNN-BiLSTM model).
Reimplemented several deep learning transcription models, for guitar tablature transcription using novel datasets.
A TINY MacOS native app to keep track of time.
A web app to rephrase text, find synonyms, and format it to be more professional, casual, or fluent.
A small web audio synthesizer that takes input from a real MIDI keyboard.
An AI research assistant that gathers, synthesizes, and cites sources on any topic, with Slack integration.
A guitar note trainer to help players learn and identify notes on the fretboard.
A live video editor with distributed rendering, file upload, and AI-generated widgets.