Thread by @chipro, When talking to people who haven’t deployed ML models, I keep hearing [...]

Chip Huyen

chipro

When talking to people who haven’t deployed ML models, I keep hearing a lot of misperceptions about ML models in production. Here are a few of them.

(1/6)

1. Deploying ML models is hard

Deploying a model for friends to play with is easy. Export trained model, create an endpoint, build a simple app. 30 mins.

Deploying it reliably is hard. Serving 1000s of requests with ms latency is hard. Keeping it up all the time is hard.

(2/6)

2. You only have a few ML models in production

Booking, eBay have 100s models in prod. Google has 10000s. An app has multiple features, each might have one or multiple models for different data slices.

You can also serve combos of several models outputs like an ensemble.

(3/6)

3. If nothing happens, model performance remains the same

ML models perform best right after training. In prod, ML systems degrade quickly bc of concept drift.

Tip: train models on data generated 6 months ago & test on current data to see how much worse they get.

(4/6)

4. You won’t need to update your models as much

One mindboggling fact about DevOps: Etsy deploys 50 times/day. Netflix 1000s times/day. AWS every 11.7 seconds.

MLOps isn’t an exemption. For online ML systems, you want to update them as fast as humanly possible.

(5/6)

Deploying ML systems isn& #39;t just about getting ML systems to the end-users.

It& #39;s about building an infrastructure so the team can be quickly alerted when something goes wrong, figure out what went wrong, test in production, roll-out/rollback updates.

It& #39;s fun!

(6/6)

You can follow @chipro.

Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: