Best Practices for Python Dependency Management

Written by Paul Kernfeld

Dependency management is like your city’s sewage system. When it’s working well, it’s easy to forget that it even exists. The only time you’ll remember it is when you experience the agony induced…

How Well Can You Predict Student Responses?

Written by Ruben Naeff

NIPS 2014 and Human Propelled Machine Learning

Last month Knewton gathered along with the global data science community at NIPS, which has become one…

Latent Skill Embedding

Written by Sid Reddy

In this post, we explore a probabilistic model of student learning and assessment that can be used to recommend personalized lesson sequences. A student using Knewton-powered products currently receives content recommendations one step at a…

How Knewton Cutover the Core of its Infrastructure from Kafka 0.7 to Kafka 0.8

Written by Christofer Hedbrandh

Kafka has been a key component of the Knewton architecture for several years now. Knewton has 17 Kafka topics consumed by 33 services. So…

Testing for Failure

Written by Philip Vegdahl

12:20am. I was jerked awake by the ringing of my phone. There were problems in production.

I logged on, desperately trying to clear the fog of sleep from my brain. There were already a handful of alerts for a variety…