concierge

A continuous learning collaborative filter¹ deployed with a light web server². Distributed updates are live (real time pubsub + delta training from model snapshots).

Live:

• redis pubsub updates, and model persistence
• using redis ordered sets to take incremental trained models and augment them with the latest events

1. using river-ml
2. using sanic

Released under the MIT License