Develop and Create DL Model for Automatic Speech Recognition (ASR) (datasets given)

The stages involved to develop that speech to text ai model for a new language are Data Preprocessing, Feature Extraction, Tokenization, Data Alignment, Model Architecture, Training, Decoding, Evaluation . I have dataset in audio and corresponding word document.
Deliverables are source code, report on the DL model steps creation with justification and also the generated model.