Dima Kniazev – Using Redis as Online Feature Store for Real-time Inference

Repo: https://github.com/Redislabs-Solution-Architects/feast-credit-scoring-demo

Description: In this talk we will explore the modern architecture for implementing real-time machine learning pipelines using Feast feature store and how Redis can be used to enable low latency prediction and scoring. We will run and look into the demo application for credit loan approval that is visualized using the Streamlit framework. Feel free to grab your laptop, clone the repository and experiment with the application. You will need Python 3.9+ and connection to Redis in-memory database: either docker version or Redis Cloud will work just fine.

Bio: Dima Kniazev is a Software Engineer at Redis. He is originally from Belarus where he got his master degree in computer science from the Belarusian State University of Informatics and Radioelectronics in 2002. He worked as a Database Developer, Data and Solutions Architect for a number of big corporations mostly here in Texas. He has a lot of experience building data pipelines for self-service BI, reporting and data science use cases. At Redis he is focusing on real-time data processing and streaming