Sifting Through The Noise of Social Media Comments With Precession

Practical lessons from building real-time topic modeling for millions of nested social media comments: three core infrastructure challenges and the technical solutions we used.

Overview

Siftree is a Chicago-based AI startup introducing advanced topic modeling algorithms to businesses and creators with active communities. Siftree invented a process to “unravel” millions of nested comments and use NLP to identify/summarize/elevate the topics of conversation in real time, with impeccable precision. Siftree partners with some of the largest agencies, artists, and creators in the world.

We want to share 3 of the largest challenges we faced when creating our infrastructure and how we overcame them.