Day 12 of 100 days of large language models

🐒 They say imitation is the sincerest form of flattery!

🍫 Edward Schmuhl created a super sweet demo for searching health supplements on user natural language queries like “top 5 helpful for painful joints”. Combines the Weaviate open source vector db, LLMs, & graphQL. Automates the graphQL query syntax including search filters derived directly from the context of the user query. It then creates product summaries of health supplements from user reviews.

🏋️ Taking Edward's demo, I stuck a 100-line, simple Streamlit UI in front of the existing FastAPI backend layer. Believe this makes it data scientist friendly & gets LLM with vector db use cases to MVP quickly for stakeholder showcasing.

Check out the Streamlit variant here
Stands on the shoulders of Edward's amazing original work here