How The Application Of Machine Learning Is Changing The Way Companies Use Data
summary
Claims that retrospective BI must give way to predict-and-prevent workflows and cites a Forbes stat: only 22% of enterprises have ML in production. Identifies concrete failure modes — brittle one‑off notebooks, missing labels, unit inconsistencies — and describes an implemented stack: checked‑in model repositories, container orchestration, API‑managed microservices and secrets, plus a probabilistic auto data wrangling layer and a self‑service deployment app (internal name Hollywood). States the trained model (not algorithm) is the moat and says this pipeline turns prototypes into repeatable production.
AI-written summary · about
chapters (15)
- 0:00Intro
- 2:00What about the Enterprise
- 3:19What is Cognos
- 5:15Analyze and fix
- 6:01Data science projects
- 6:54No shortage of use cases
- 8:39No selfservice AI
- 9:43Oneoff projects
- 10:09Data wrangling
- 11:10Development best practices
- 11:46Supportability
- 12:50Architecture
- 13:58Pilot
- 15:06Summary
- 16:32Questions
related talks
27mA Look Inside Applied Machine Learning
30mThe State of Production Machine Learning in 2023
30mChallenges of machine learning: The new ML lifecycle
27mKeys to Building Machine Learning Systems
06mMachine learning is about your data and deployment
50mRealities & Misconceptions about Enterprise Applications of AI/ML and Data
19mSustaining ML in the enterprise
52mThe ML Lifecycle in Combat Conditions