Artificial Intelligence Intern – Audio Data Classification
This experience involved developing a CNN-based music instrument recognition system, annotating audio data with Mel-spectrogram features to train and evaluate the deep learning model. I was responsible for labeling the audio data for instrument types, ensuring that each training sample was correctly tagged for model performance. The role encompassed preparing audio data, transforming it into features, and assigning accurate class labels for supervised classification. • Processed and labeled audio files with corresponding instrument classes. • Integrated Mel-spectrogram representations using TensorFlow for labeling. • Built and managed annotation pipeline for audio input data. • Contributed to training deep learning models for multi-class classification.