Audio & Speech AI
From waveforms to Whisper
How machines hear and speak — spectrograms, speech recognition, text-to-speech, and the audio models behind voice assistants.
Intermediate3 modules · 10 lessons~82 min775 XPAligned with Hugging Face Audio CourseStart with: Deep Learning
Progress0/10
Module
Audio as Data
Waveforms, spectrograms, and teaching networks to hear
Sound Is a Waveform8 min · 70 XPSound Is a Waveform
Spectrograms: Sound as an Image8 min · 75 XPSpectrograms: Sound as an Image
Audio Classification8 min · 75 XPAudio Classification
Module
Speech Recognition & Synthesis
Machines that transcribe, translate, and talk
The ASR Pipeline9 min · 80 XPThe ASR Pipeline
Whisper: One Model, Many Languages9 min · 85 XPWhisper: One Model, Many Languages
Text-to-Speech8 min · 80 XPText-to-Speech
Voice Cloning & Its Ethics8 min · 75 XPVoice Cloning & Its Ethics
Module
Music & Audio Generation
When the model composes
Why Generating Audio Is Hard8 min · 80 XPWhy Generating Audio Is Hard
Modern Audio Transformers9 min · 85 XPModern Audio Transformers
Audio AI in the Wild7 min · 70 XPAudio AI in the Wild
Capstone project
Train an ear
Build a sound classifier or speech experiment — classify genres, detect keywords, or analyze spectrograms. The feature extraction choices are the interesting part.
Deliverable: A notebook with spectrogram visualizations and a working classifier on real audio.
Milestones
Finish every lesson to earn the Audio & Speech AI Certificate.