8000 GitHub - chaba-victor/Real-time-data-pipelin
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

chaba-victor/Real-time-data-pipelin

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Building a production-ready feature pipeline is real-time

Tools

Table of contents

Aim

The ultimate goal is to build a trading boat that is powered by ML.

Before even thinking about how the ML model will do its thing, we'll need to design, develop and deploy a real-time feature pipeline that produces the features needed by the model both at training and at inference.

The pipeline has 3 parts:

  • Ingestion of raw data from an external service. This would be raw trades. Kraken Websocket API will do.

  • Transform these trades into features for the ML model.

  • Saving these features in a Feature Store, to be fetched by the ML modelto generate both the training data and real-time predictions.

In a real-world setting, each of the above processes is implemented as a separate service, and communication between these services happens through a message broker like Kafka.

This way, yoursystem becomes scalable by spinning up more containers as needed, and leveraging Kafka consumer groups.

Wanna learn more about ML?

→ Take a look 🤗

About

No description, website, or topics provided.

Resources

48C8

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published
0