AI for the Practical AI Podcast

Background:

Practical AI (https://practicalai.fm/) is one of the most popular AI podcasts. As of now the publishing of the podcast on the main website is fairly bland with episode listings and basic information. Given the topic of the podcast and the value in the backlog of 300+ episodes, the hosts of the podcast would like to create an AI-driven website for the podcast which can dynamically engage listeners with the content and drive further value for listeners, thus increasing engagement.

Project Description:

In this project, students will utilize the Prediction Guard GenAI platform (which has LLMs, vision models, embeddings, transcription, etc.) to: (1) create a chatbot in which listeners and those curious about AI can ask questions, get answers based on episode content, and be dynamically linked into relevant episodes; (2) create AI-driven automations to process show requests, analyze news trends for potential show guests, and generally improve operations; and (3) create AI-driven automations that will draft social media content linking existing episodes with trending AI-related news.

Outcomes:

1. A RAG chatbot system backed by all existing Practical AI episodes and capable of being updated as episodes are published

2. A draft website or embeddable widgets for automating show requests and improving operations

3. A pipeline for automating draft social media content linking to relevant podcast episodes