Building ML and AI apps with Shiny for Python
summary
An assistant named Shiny Assistant generates runnable Shiny apps in R or Python using Claude 3.5 Sonnet and loads them into Shinylive so code runs in-browser; it builds a web-API client for embedding.io, assembles chunks for RAG, and calls OpenAI via raw cURL to produce answers. It fixes issues like missing magrittr or dplyr when prompted, warns that not all packages compile to WebAssembly, and often produces usable scaffolds in about five minutes.
AI-written summary · about
chapters (11)
- 0:00Shiny Assistant intro
- 1:07Live demo of Shiny Assistant
- 2:41Debugging generated code
- 5:51AI output formatting
- 7:42Shinylive secure execution
- 9:49Python app with API
- 13:03RAG app with OpenAI
- 16:20Shiny Assistant limits
- 17:45Real-world examples
- 18:48Future of Shiny Assistant
- 19:11Q&A on LLMs
related talks
18mAI and Shiny for Python: Unlocking New Possibilities
20mOld Apps, New Tricks: How AI can write Automated Tests for your Shiny Apps
20mWeb applications with Shiny and React and AI
36mGenerative AI + Python: Unlocking Efficiency, Personalization, and Insight
45mPositron AI Session
40mCreating Interactive GenAI applications powered by Streamlit, open source LLMs
52mBuild an Intelligent Python App, No Infra Hassles!
59mMaking the most of artificial and human intelligence for data science