Machine learning and artificial intelligence are all the rage these days, with companies across every industry trying to use them within their software systems.

When used properly, machine learning can bring benefits over traditional software systems but it is important to understand the potential costs of implementing such a system as well.

This video is based on a paper published in 2015 by researchers at Google and highlights some potential gotchas related to incurring technical debt when building machine learning models.

Link to paper -- "Hidden Technical Debt in Machine Learning Systems": https://papers.nips.cc/paper/5656-hid...

---

Website:
https://devopsdirective.com/

---

Timestamps:
0:00 - Intro
0:56 - Technical debt overview
1:33 - Complex models erode boundaries
2:50 - Data dependencies cost more than code dependencies
4:53 - ML System antipatterns
6:02 - Fighting technical debt in ML systems
7:14 - Outro

---
Join the Community:
💬 Discord:   / discord  
💻 GitHub: https://github.com/sidpalas/devops-di...
🐥 Twitter:   / sidpalas  
👨‍💼 LinkedIn:   / sid-palas  
🌐 Website: https://devopsdirective.com

---
Community Size at Time of Posting:
Subscribers: 104
Channel Views: 2013