Problem Description

Looking to implement the Python Whisper JAX API to translate + transcribe large audio files and youtube videos in bulk on Google Cloud Platform

This is the repo, please read before applying:
https://github.com/sanchit-gandhi/whisper-jax

Acceptance Criteria

Reference example here: https://huggingface.co/spaces/sanchit-gandhi/whisper-jax

I need an API ONLY, no UI required.

It should accept a URL, and be able to queue or process multiple jobs simultaneously and receive corresponding translations + transcriptions for each.
Transcription in original language | Transcription in English | timestamps

Technical Details

Needs to be hosted on a GCP container, GPU/TMU

Inputs:
url of either an audio or video file {string}
Include Timestamps: {bool}
Callback url of a webhook {string}

On success/error it should call the supplied webhook with the transcription or error in the body.

Needs to be scalable for a production environment.