Mikiko

📝 What kind of MLOps team are you? [Part2/3]
#mlops #productionml #dataops #mlsystems

🔍 Zeroing in on the ones that oftentimes constitute the ML Org or the Data org:

⛑ Enabling teams - Help the DS & Product folks get those models out the door using the internal plateforms & capabilities provided by the CST

⚙️ Complicated Subsystem team - Focused on maintaining & expanding the extremely technical solution they own

👷🏻‍♀️The Platform Team - Owns unified & integrated experience.

Apr 18, 2023 at 3:00pmWeb
Show threadJP de VooghtApr 18, 2023
@mikiko great question Mikiko! I am thinking 3 types could come into play e.g. one providing a delightful experience with data & ML tools to stream-aligned teams, one enabling the SA teams with knowledge about the tools and one complicated sub-system digging into a particularly large (e.g. probabilistic graphical) model too large to be handled by an SA team, even empowered.
Show threadMikikoApr 18, 2023

@jdevoo
I've talked to a few folks about their perspectives and the common theme seems to be "pick 2 of 3".

For example over at Riot games:

"Pretty interesting idea. I'm not super sure where leauge lands - i guess we're closer to only enablement and platform integration as a shared responsibility."

Another person said:
"This is true from what I've seen too. My team does ⛑ and👷🏻‍♀️ and there's other teams that do ⚙️ and👷🏻‍♀️"

Show threadJP de VooghtApr 18, 2023
@mikiko at least 👷🏻‍♀️ in both cases - so there's perhaps less questioning of the purpose itself as raised by Sam here? https://samnewman.io/blog/2023/02/08/dont-call-it-a-platform/
Sam Newman - Don't Call It A Platform

Show threadMikikoApr 19, 2023
@jdevoo Agreed! And I loved Sam's post, have referenced it several times in convo with folks! 😃