What Data Scientists Need to Know about MLOps Principles

Podcast interview of MLOps engineer Mikiko Baseley hosted by Daliana Liu

Original Blog Post: What Data Scientists Need to Know about MLOps Principles
Authors: Daliana Liu (host) and Mikiko Baseley (guest)
Published: 2023-08-31

My summary:

First, a definition: What is an MLOps engineer?

In short, an MLOps engineer is responsible for making data science / ML models work in production, and maintaining and monitoring these production models.

Mikiko Baseley transitioned from data scientist to MLOps engineer, so I think her perspective is really useful for data scientists to hear, especially if we have models that we want to move into production.

One point I want to highlight is from her point of view as an MLOps Engineer, Mikiko lists 4 things that all data scientists should try to know and understand [at 29:49]:

1. Version control (i.e. git)
2. Containerization and packaging strategies
3. Basic web technology and deployment patterns
4. Testing

Here is the link if the embed above doesn’t work: Link to Podcast on Spotify