Vibe Coding with Prediction Guard

Background:

Prediction Guard’s customers are using the platform to deploy private, secure GenAI solutions within regulated industries or in scenarios with confidential data. Often these customers want to use the best-in-class AI tools on top of their private AI infrastructure (Prediction Guard). This means that, in addition to custom development, they want to use best-in-class AI code assistants plugged into private AI models, such that their code, config, data, etc. is not leaked outside of their infrastructure or network.

Project Description:

This project will extend an existing “vibe coding” platform (such as Open Hands) to work flawlessly with private models running Prediction Guard. The extension may involve contributions directly to the Open Hands (or other relevant) project. Once integration is achieved, students will utilize the vibe coding platform, along with starter kit code from Prediction Guard, to create guides for Prediction Guard’s customers for rapid development of RAG chatbots, analytics systems, agents, etc. These guides will be published along with the integration to help Prediction Guard customers achieve rapid time-to-value.

Outcomes:

1. Integration code, config and docs for a relevant vibe coding platform such as All Hands/ Open Hands

2. 3 guides for using the vibe coding platform (along with starter kit code from PG) to create relevant full applications for a RAG chatbot, analytics system, and AI agent