TechPrufer
Get started
Slack
Slack logoSlack · Feature trail

Recommend API

1 post · Sep 2022

What this trail covers

This thread tracks the evolution of Slack's recommendation capabilities, focusing on the development of a unified Recommend API. Initial posts detail the creation of this API, which provides a common framework for building and deploying various recommendation use cases across the product. The API reuses infrastructure for data processing, model training, candidate generation, and monitoring, enabling faster bootstrapping of new recommendation features. Subsequent posts will detail specific recommendation models and their impact.

All Slack trails
  1. Recommend APIThis post introduces the Recommend API, a unified framework for building and deploying recommendation use cases at Slack. It details the API's architecture, which reuses common infrastructure for data processing, model training, candidate generation, and monitoring. The post outlines how the API enables rapid development of new recommendation features, providing examples of its application in Composer DMs, creator invite flows, Slackbot channel suggestions, channel browser recommendations, and invite reminder send-time optimization. It also describes the backend workflow, data processing pipelines, model training pipelines, and data privacy considerations.