Get the latest tech news

How Novy Built Zelma for Emily Oster Using OpenAI and Postgres

Are you thinking about building a polished AI powered app to explore a large data set? Worried about how to handle data accuracy issues? Do you want to know the pro's and con's of using pgvector with Postgres over a dedicated vector store? Are you interested in hearing tips and tricks for reducing LLM latency? If so then this interview with Luke Van Seters about how he and his team built Zelma is for you! This interview has lots of juicy technical details learnt from shipping a production AI app

This interview has lots of juicy technical details learnt from shipping a production AI application used by thousands of parents across America. Zelma was developed with Novy in conjunction with Emily Oster an economist at Brown University and a bestselling author known for her work in parenting. 24 hour prototype of productPolished version of this screen that shipped Accuracy Issues: Data partitioning and denormalization led to incorrect results.

Get the Android app