Client - Pfizer

Implemented an emotional text-to-speech system that aided 1000+ users in pronouncing English text along with emotion. In addition, I integrated pre-trained neural network models (Tacotron and an Emotion-Mapping Model) into a Flask backend pipeline. Ultimately, I helped design an interactive user interface that enables users to input text and choose from 5 emotions, generating downloadable audio output.

Revamped the communication infrastructure between the speech training application and backend by replacing socket communication with REST APIs. Moreover, I constructed and managed a MySQL database for efficient data storage and retrieval. I worked on this project in the Speech Processing Lab over the course of a year.