Continuous Delivery Foundation MLOps Special Interest Group, 15 Jul 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: CDF - SIG MLOps Meeting 2021-07-15

Description

For more Continuous Delivery Foundation content, check out our blog: https://cd.foundation/blog/

A

A

Hi there hey hello.

A

So have you uh have you been along to uh one of these sessions before? Is this your first time.

B

um This is my first time.

A

Okay, um so I'm terry, I I chair the um emelops roadmap work.

B

um I'll also introduce myself sorry can't currently.

A

Sorry can can, can you hear me.

B

Yeah, I can hear you now.

A

B

A

I I chair the uh ml ops roadmap work.

A

A

B

um I think it might be my internet because you are somehow breaking off uh after you said you shared the mlops roadmap.

B

Give me give me a moment to reconnect sorry.

B

B

B

um I'm not sure, probably you're having some internet issues, maybe.

A

A

Hello should this should be a bit better. Now: hey hello, yeah, it's better! Now, yeah! It's like I've got a hardware problem on another studio, so I've switched locations um yeah, so uh apologies. uh I was just trying to uh do some introductions. um Yeah.

A

What how have you found the the the road map? What's brought you along today.

B

Yeah um so basically I was exploring um some of the.

B

Cd foundations channels- and lately I had been working um some some time back. I had been working on data editops and some of the activities with a few projects.

B

So uh during the time some of the pipelines were really sort of a linear thing and they do general sort of regressions and all, but there were also ml models and all and uh different data team used to work on that. So I got interested and wanted to find out what's happening in this space.

B

So that's that's. Why.

A

I'm here well what what we're trying to do with the with the road map is: um take a look at the whole problem, space for machine learning in production environments uh and create a resource that allows everyone to to get a really good overview of all of the challenges that you're likely to face.

A

When you move into that that problem domain.

A

What we're seeing at the moment is lots of people working in machine learning, applications.

A

But they're being quite a narrow perspective, and so there's a there's, a big problem with people not knowing what they don't know and and then getting into difficulties when they try and put their models into production, because there were a lot of things that uh would have been easier to tackle. Had they been thinking about them earlier in the process.

A

And then the other thing that we're trying to do is to promote some collaboration about uh understanding these problems at a sort of vendor and tooling level, so that anyone who's providing tooling for mlrs will have a shared understanding of what the problems are, that their customers are actually facing.

A

uh Try and get people to implement good solutions to all of those problems, so that we've got tools. We need to be able to solve the problems. We've actually got.

A

So um have you had a read through the um the roadmap document.

B

No not yet, where can I look at the document.

A

So uh in the um in the cd foundation, github repository uh you'll you'll find the the sig ml ops it repo and if you drill down into there, you'll find uh the uh annual release documents for the road map.

A

So that's a good place to start. If you want to get up to speed with you know where we are right now and what we think the main challenges are.

A

Yeah, so if, if you, uh if you haven't read through the 2021 document um and if there's anything in there, that you think we've missed uh or that needs expanding upon, then absolutely feel free to either give me a shout or just submit an edit as a as a pull request on the document and we'll review it and include it.

A

The idea is this: is a uh a collaborative group um we're a little bit quiet at the moment because everyone's taking the holidays right now, uh but usually there's some discussion going on in this forum and and then, if you're not already joined, you know, join the cdf slack group and come into the uh cml ops working area um and you know feel free to join the conversations there.

A

B

You for showing the resources where they are.

A

No problem is there anything I can help you with today.

A

B

It's better, I just go through dogs. First.

A

Okay, well, uh hopefully, you find it useful uh look forward to seeing you at future sessions.

B

B