Streamlined DevOps and AI Adoption with Gal Marder at JFrog swampUP 2024

This is Techron tv. Hey, everyone. We're back here at Swamp Up in Austin.

Uh, it's day two, it's the afternoon, and now we're getting to some of the Jfr executives I, that I know for years. But I'd like to have them on the show. We'll get a little bit more insight and try to give, for those of you who couldn't make it out here, to swap up a sense of what's been going on, what the sessions have been like, maybe you can make a plans for next year's Swamp Up.

And next guest is my friend, Gal Marder. Gal is VP of Strategy. VP of Strategy.

Yeah. VP of Strategy at at Jfr. If you've watched Techstrong tv, you've seen GAL on before.

Um, gal welcome. Thank you, Alan. Thanks for having me.

First Of all, congratulations. It's been a great swamp up. I love the energy, love the, the passion as always.

But it's also been chockfull of announcements and, and news, which for us is good. So, you know, going over the big news, um, uh, the Nvidia partnership, uh, GitHub partnership, uh, there's been some security news and so forth, but your role in VP of of strategy really revolves around a lot of these partnerships. Yeah.

Strategic partnerships. Correct. If you don't mind, I'd like to dive in a little bit to the NVIDIA partnership, but then talk kind of a broader level, the kinds of things you are looking to accomplish with partners Hardships, Of course.

So, so first about sw, I cannot just disregard it. Uh, look, for us, swamp Up is more than an event. It's a heartbeat for the year.

And coming here and getting the energy from, from our, our customers and the Love and seeing the Frog so excited. It, it, it kind of every year surprised me again. Yeah, No, it Does.

But, but that is how it feels and it's very valuable for us getting the feedback, getting the energy. Uh, it's amazing. So happy to be here again, obviously, uh, this year, and happy to have you with us at Swamp Up Always.

Uh, specifically regarding, uh, the ecosystem and partnership and, and how, how I look at it and what we did this this year. So going back to the history of J Rog, we were always part of the ecosystem and we couldn't, we couldn't grow without it, you know, back in the days it was, uh, with Jenkins, CloudBees and, and Artifactory. This is, this is how we started.

And sure. We have this phrase saying too integrated to fail. And, and I truly believe it.

And, and as our industry evolves and matures, this becomes even more critical because when we look at the responsibilities and, and what developers and engineers are being required to take care of every year, this is getting more, more and more complicated. Uh, you know, DevOps started with speed and automation, but then security came in. Yes.

And then there's more responsibility that has to do with compliance as well. And now we have AI and MLOps is, is knocking on our doors as well. And it's our commitment and our duty to provide a holistic solution that will make it bearable.

So we make, it will keep us sufficient. And in that kind of environment, there's no chance that one vendor can do that all by themselves. No.

Okay. Not in today's world. Not in today's world.

Exactly. Uh, so the way to achieve it is, is to be, to integrate, to fail. Sure.

Is to find those partners that are playing in the ecosystem, that share the values with you, that provides, uh, high, high quality solutions and, and serve the community and listening to the community and the customers, and work together with them, uh, to build the full solution. And I think this year, uh, we've been doing good at that, at that front. And as you mentioned, we partnered with GitHub.

Yep. Uh, which we see in the industry a lot as kind of the, uh, the leading for everything that has to do with store centric, right. Uh, activities around DevOps.

And we're obviously focused on, on, on binaries, and we see a lot of our customers using both. They came to us and say, guys, make it easier for us. Right.

Go together, not only make it easier for us, but also provide us with the synergy and more value that you can provide each alone. And I'm very happy for our partnership there. And, and, uh, and Thomas and, and the GitHub team, were haing and we're working together.

And, and this is, uh, kind of again another announcement here of, um, co-pilot, uh, extension with, uh, with our capabilities. It was very exciting. And on the ML lops front, uh, and, and AI front, we're very happy to partner obviously with, uh, with Nvidia.

We can talk about the details, uh, in a second, but making it easier for, to consume and bring AI to production. That's the gist of it. Yeah.

I do want to jump in on that and then we can go back out to some of the other partnerships. You know, some people look at it and say, Nvidia, they make chips. What is the relationship with jfr?

But what I think people who aren't in the know, don't know, 'cause they're not in the know is Nvidia is developing a whole software ecosystem around the use of their chips. Because, you know, hardware is a funny thing today. You've got the best hardware, but it doesn't take long for the rest of the industry to not copy, but to catch up Yep.

On your designs, on your, you know, on your, your speeds and feeds. But ultimately, software prevails, right? You Correct.

So if you can combine or surround your hardware with software that you can control, it's a lot harder for your competitors to, to take that space. And so that's kind of NVIDIA's play, right? So I'm imagining the Jfr integration is with the Nvidia and a lot of it's open source too, in the case of Nvidia.

But I, you know, I, I, that's where that's happening. Can you go into detail a little bit? Of Course.

Of course. So, so look, I'm, I'm obviously not leading the strategy of Nvidia, but, but what do you say makes sense? Yeah.

And, and that's probably the case. And, and Nvidia has been doing software for, uh, for years mostly to enable and make it easier to kind of take advantage of, of hardware. What Yeah.

Yeah. We could look at Cuda and Triton Yes. And many things that they have done, uh, in the past.

And I think that the current challenge that we see with, uh, with AI and LLMs is that it's very promising. It's amazing. The opportunities are endless, but when you wanna take some of it, take an open source model, put it in production, uh, at scale, it's very hard.

Yes. And although NVIDIA provides obviously the best GPUs and amazing hardware, tuning something or taking a model, uh, to make sure it takes advantage of this hardware the right way is not easy. Yeah.

And, and Nvidia came, came with, with an, an amazing solution for that will say, look, here's what we'll do for you. We will take these very popular open source models, the leading ones. Okay.

And we'll make sure to do all the work that needs to be done in order for them to take advantage of our hardware in the best way. And to make it easy for our customers. We'll just take all of this, put it in a docker.

Okay. And this docker will be very easy to consume and put on our hardware, and the magic will happen. And I think it's a super smart, smart move.

They call it, uh, nims. Yeah. Okay.

And video inference microservices. That, that's what makes stands for. And, uh, and, and this is, there's a brilliant move, but the challenge is that when you want to take this great magic image, uh, magic name, that, that knows how to take advantage of the hardware and want to take it into the organization and put it as part of your application, you need to embed it into your software supply chain.

You need to somehow curate that into the organization. You need to keep a copy of it. You need to, to version it.

You need to do your tests together with your applications. Synchronize the other dependencies. Software is software.

Software is software. And I think this is where the synergy comes in. When Nvidia did a great job with, with memes, now embed it into your software supply chain.

Taking advantage of the existing policies and pro process you also already implemented with the jfr platform within the organization and kind of make NIM fit into that naturally makes it much easier for enterprises to adopt, uh, Nim and kind of creates an ease of, and a tremendous ease of use to kind of get AI to production while keeping your existing policies and processes in place and leveraging. I love it. Now, my other impression in listening that this is, this is the first pass, if you will.

This is the first layer, right? It's like peeling an onion. Correct.

The layers of the onion. Don't say anything you're not supposed to announce, but will we see perhaps other layers to this onion Yes. Over the coming weeks or months?

Yeah, So working on a, we're working on a full roadmap with, uh, uh, with Nvidia in general. I believe that when you have the right partners, uh, when you have the right collaborations, uh, this is where it's not a one time thing. You, you're planning a roadmap.

It's, it's it's software, you know? Yeah, yeah, yeah. It's evolving.

Uh, so we're definitely, uh, gonna see more things, uh, coming from this, uh, collaboration. I cannot, uh, I cannot elaborate much more than that, but I can say that it's gonna be all about how we make it easy to, um, uh, build enterprise grade responsible ai, uh, with everything that Nvidia can contribute to that. And can jfo can contribute to that, and you'll have to stay tuned to kind of get On that.

Might there be other partners in that? Well, Yes, there might be other partners that well, but specifically to, to this partnership, uh, uh, we're, we're aiming for that as well. As we know, we, earlier this year, we also acquired QAC that is aiming to, to, uh, to achieve the same, and again, with QAC as well.

You know, they have a very, or we have now with J four ml a very comprehensive, uh, uh, platform. Uh, but it will never cover everything. Yeah.

So we definitely plan to add on more partnership and more capabilities and integrate them into jfr ML and the ML platform in general to make sure that, uh, producing AI becomes, uh, easier for engineers around enterprises. So gal, before I got involved in, you know, media and stuff, I helped start several companies and I was always the VP of biz dev corp, dev, SVP, what have you, strategy chief strategy officer. So I have a little experience in partnerships and strategic relationships.

Would you say most, most companies come to you to Jfr saying, Hey, we'd like to work with you? Or are you specifically targeting outbound or a little of both? It's a great question.

So, I, I'm, you know, humbly say in that, in our position today as the single source of truth and system of records of your software supply chain, we have a lot of vendors coming to us to partner, partner with us, uh, because we provide kind of a common layer for things. So, so I, it Well, it's part of being a platform. Exactly.

It's part of being a platform and part, part of being the, the system of record for, uh, for a domain and, and, and part of us being a significant player in this ecosystem, uh, uh, for a while, this is great. But, but I must say that, that I'm constantly challenging myself with being intentional about things, with saying, okay, obviously I listen to what partners are coming and, and we're collaborating with everyone and making sure that, uh, again, that, that we're easy to use. Uh, but we also wanna make sure that it aligns with, with our strategy.

Uh, and sometimes there we're reaching out as well. So it's a little of both. Okay.

Probably in terms of numbers more, more reaching out to us than Yeah. We're reaching out to them. But it's a combination of both.

And It's a two-way Street. It's a two-way street for sure. And we're an open platform.

We love that everyone partners with us, obviously, like everything you have to set your priorities and say where you invest more. And we try to be very intentional about that and, and making sure that the partnership we invest, we invest the most in are aligned with the strategy and with the value that, uh, that the customers are, are getting. I think that this is the, the, the first thing, okay, if we create a lot of value for our customers, it's a good thing.

Everything will, Everything will follow. Right. Everything Will Follow.

Because that's the reason to do these. You know, again, over the years I did a lot of deals. Gee whiz, you know, it seemed cool, it was cool technology, but if there's not a, a customer at the end of the stream that says, this is valuable to me.

Yeah. It's for nothing. Exactly.

Anyway, thank you so much. Thank you all. It was great.

Being Here is great. We're gonna take some pictures. We're live, well, we're not live, but we are in Austin at the 10th annual Jfr Swamp Up.

Thank you. We'll be back in a moment. Actually, we've got some special guests coming up.