Project Goal

Build a simple MVP of our offline, privacy-first AI platform that allows users to upload documents and receive AI-generated summaries or classifications locally. This prototype will serve as the foundation for our future edge-AI product.

Main Objective / Tasks & Deliverables

Learners will:

• Set up a lightweight local backend using Python (e.g., FastAPI or Flask)
• Connect the backend to a quantized open-source language model (e.g., via llama.cpp or GGUF format)
• Develop a simple front-end interface or CLI for uploading documents and displaying AI outputs
• Implement basic support for common file types (PDF, .txt, .docx)
• Package the MVP for local testing/use on a laptop or NUC
• Document their work and provide setup instructions for reproducibility

Deliverables:

• A working MVP that runs entirely offline on local hardware
• Source code and setup documentation
• Basic UI or command-line interface for demo purposes

Project Outcomes and Deliverables

Learners are expected to complete the following:

1. Local Backend/API

• A lightweight backend (e.g., using FastAPI or Flask) that can run locally.
• Routes for uploading documents and returning AI-generated summaries or tags.

1. Model Integration

• Integration of an open-source, quantized LLM (e.g., via llama.cpp or GGUF format).
• Ensure the model runs entirely offline on local hardware.

1. User Interface (UI or CLI)

• A basic interface (web dashboard or command-line tool) that allows users to:
• Upload a text document (PDF, .txt, .docx)
• View the summarized/classified output

1. Documentation

• Clear setup instructions to reproduce the environment locally.
• Short user guide explaining how to use the MVP.

1. Demo

• A working demonstration (recorded or live) of the MVP running locally.