Modal Development

We have a python app that uses GPT API running on a local machine using local GPUs. The front end is single screen PyQT5. We now need to scale this to a web service running on modal.com, converting the PyQT5 UI to Gradio for the UI/orchestration layer so we can run batches on multiple instances and easily scale GPU and CPU resources as needed. We've scoped this initial deployment at about 80 hours for a developer with these proven skills (modal and gradio), with future work to continue based on performance. Our app batch uses batch downloaded docs. via API, parses and uses LLM to generate a docx report and set of .csv files. Analyzed content is based in the biosciences area and in future phase we'll be adding functionality to analyze bio-sequences using Blast (supported on modal.com). Experienced developer will already know: Modal enables:
Full serverless execution
Define container environments
Attach GPUs with a single line of code.
Serve functions as web endpoints.
Deploy and monitor persistent scheduled jobs.
Modal requires zero configuration- everything is code (no YAML)