Overview
LLMs can help chatbots offer users more personalized and responsive conversations. However, this flexibility also causes LLMs to make things up and possibly even say harmful things. While chatting with an LLM is simple, integrating one into a chatbot can be daunting. Delvin aims to provide tools for people to build safe, grounded conversations for their target audience.
In this project, I had the opportunity to work on the back end and the front end, learning more about how these systems interact. Additionally, I worked on a retrieval augmented generation (RAG) system for the service, experimenting with different settings, tools, and techniques to boost accuracy.
My Roles | Our Tools |
---|---|
Full-stack developer UX/UI Designer | Python Django LlamaIndex TypeScript React Shadcn Docusaurus Celery Supabase |
Check out my work!
Code | Design |
---|---|
Merged work | Guided Conversations Subscription Tier System |