- Published on
Just Enough Software Engineering (For Data Science)
Well, here's the slides from a presentation I gave to a group of SMU students a few months ago. There's a lot of technical (over)simplification here because the audience doesn't really care about the finer points of software architecture, they just want to make a model more useful to the other developers on their team. Tony might post a (much better informed) rebuttal below, so I'll tell you to take all of this with a grain of salt. I'm no expert, I'm just okay at getting us started with an ML driven project.
Probably the most important part of this presentation is slide 4 - Know where your model fits.
- What are your inputs?
- What are your outputs?
- What are you actually predicting?
After that, knowing how to hand it to other developers is super helpful!
Here's some sensible defaults for shipping ML projects.
So, go answer those three questions up top and try to make something. Learn a little about REST APIs, FastAPI, and Docker and you'll be a way stronger contributor on your team.