Cloud Infrastructure and MLOps with Deepa Mani at JFrog swampUP 2024

This is Textron tv. Hi everyone. Alan Shimel back here at Swamp Up, wrapping up our day two coverage.

Um, I'm really ha happy to have this next person with us. She's from Informatica. Yes.

Her name is Deepa Mani. Yes, Dipa. Welcome to Text Drunk tv.

How are you? Thank You. I'm doing good.

How are you? I'm doing really well. Mm-Hmm.

So we'll get into more stuff, but let's start Deepa. What is your role in Informatica? Oh, I am, uh, principal DevOps engineer.

So I specialize in, uh, managing the cloud infrastructure, artifactory management. So recently immersing myself into the practical aspects of, uh, MLOps. So this is what I do.

MLOps. Yeah. Well, wow.

You're a rock star. Thank you. So I think most of our audience knows who Informatica is, but there might be some people who don't.

But how would you describe Informatica to them? I would say Informatica is a leader in, uh, data management solutions. Now with our, our very own cloud G pt.

It's a AI power data management, uh, solutions. So it's a cloud-based management platform, uh, that helps, uh, business to, uh, process and manage large amount of, uh, data. So is it it's data storage database?

We manage all of the above. All Of the above, Yeah. Okay.

And I mean, we're here at Jfr Swamp Up. I'm, and you mentioned you're an Artifactory manager, so you're using Jfr? Yes.

Why? Why is, uh, we want to have a centralized artifactory management, so, which is highly scalable, you know, which is a high performance so that it ensures our developers or DevOps, uh, team to store manage version the binary artifacts and packages throughout their, uh, development and deployment life cycles. So that's why we want to have understood, If, you know, how long have you been using Artifactory for this?

I think it's been quite a long time. I would say five, Five years maybe. Maybe That's a long time.

Yeah. And again, I don't know if you were there then and you would know, but what, what, what did you do before Artifactory? So before Artifactory, uh, I would say we were pulling, basically, as I mentioned, you know, like Informatica is about, right now, we are fully into cloud, got so multiple, uh, container images or whatever from the public repositories.

So we don't want to have that. So that's the reason, you know, like, uh, we want to have one SaaS solution to maintain the infrastructure and whatnot to us. So all we do is, you know, just bump in our artifacts into Jfr, so, which the DevOps or the development team use throughout their lifecycle.

So That makes it a lot easier. So I've spoken to many customers of Jfr and there's a, there's sort of a path, right? They start without a artifactory just like you Mm-Hmm.

And then they say, you know, now that we have everything in a repo like that, all our artifacts and binaries, the next thing I worry about is the security of them. Oh yeah. Yes.

That's also one of the reasons. Uh, yeah. And we, we should start using X-ray.

Yeah. We are already Using it also using X-rays. Well, yeah.

Using x-ray. So you, so now not only are you just storing the artifacts and binaries, but you're scanning them for security? Yes, yes, we do.

So to highly maintain our, you know, the releases, so we use x-ray, uh, as well as part of our, Yeah. And again, if you can talk about a find, if you can't, we understand. Are there other jfr solutions that you're using or looking at?

Oh Yeah. So we are using, uh, jfr self-hosted edges. What I mean by it is that we have customers spread across globally.

So with the self-hosted edges, it's like a caching mechanism. So, which enables our customers to pull the artifacts real quick, which reduces the latency and high performance reduces Bandwidth and everything else. Yeah, yeah.

And also scalability. So it is important for us, uh, to bring up to spun up our infrastructure real very quick. So with these edges, it helps us to, you know, just bring up our infrastructure in like few minutes so that, you know, we don't have to wait for the images to, you know, our pods or deployments to kick in.

So, yeah. Excellent. Now you mentioned MLOps.

Yeah. A lot of people talking about ai MLOps. Again, if you can talk about it, but if not, it's okay.

What exactly are, are you using the new jfr ml or what are you doing with MLOps? Right now we are not using jfr ml, but we are, you know, like trying to use jfr ml, you know, like we also recently had our, you know, conversations with. Right.

Well, for your use case it would probably help. Yes. Right?

Yes. Yes. Sounds like it.

Um, and overall, I mean, level of satisfaction with, with the JFR products has it? Oh yeah, it is. It is.

So we are, that's why you, so I would say, you know, we are one of those customers. We are like highly dependent on jfr for our Artifactory essay, for our X-rays. And uh, you know, I also feel the customer support from Jfr, you know, if any issues, you know, it's the Re No, that's very quick.

Yeah. And, and you know, I think it manifests itself here at this event. Mm-Hmm.

The focus on customer Yes. Is, is kind of unique for them. What about, is this your first swamp up?

Uh, no, this is my second swamp up. So you were in San Jose last year? Yes, I'm in Jose last Year.

What'd you think? Yeah, it's good. So last year it was more about DevOps DevSecOps.

Yeah. So this year it's more into mops, which I'm interested and everyone are interested with. What about ai?

Oh yeah, it is good. So all the sessions were really very good. So more informative as well.

You know, we can get a few ideas and then we can see how we can implement in our use case. So that Session, do you think Gen AI is going to be a big, uh, Of course Thing for, for specifically for Informatica? Yeah, it is.

So we also have our own, uh, GPT Al already, so Oh, really? It's, yeah. Yeah.

So very cool. Yeah, we have our Clara GPT, so it's going to be, you know, like Jenna is always, it's the, it's the future. Yeah, Absolutely.

Yeah. Well, I want to thank you for coming on. Thank you.

This was really easy. I know you were a little, a little nervous. Yeah.

But you did fine. You, this was a great conversation. Thank You.

Continued success and good luck with with Thank you. The Informatica. Yeah.

Thank you So much.