Johan van Amersfoort

Johan van Amersfoort: [00:00:00] So looking at one of the movements is a dead lift. Yep. What if my deadlift is 280 pounds? Yeah. And I want to improve on certain aspects of it. So I measure other weights I can lift, like a Romanian deadlift has a certain weight. And based on those numbers, I can adjust my training to increase or to get it more stable or whatsoever.

And AI is a perfect solution for that. Yeah, literally within a day I had it ready. Oh, wow. And just to use a standardized model. Yeah. I used a coding assistant to build the application for me on a kind of per function basis. Yeah. And now it's, it's ready. The best thing of. All it runs on a NVIDIA Jetson device.

Oh, so it's a tiny device like a Raspberry Pie. Yeah. Okay. Basically on Steroids with a GPU. Oh wow.

Ashish Rajan: Hello. Welcome to another episode of Cloud Security Podcast. I've got Johan with me, man thanks for coming in.

Johan van Amersfoort: Yeah. Thanks for having me.

Ashish Rajan: Just to get set some context for people, could you share a bit about your background?

How did you get stuck into this AI container world?

Johan van Amersfoort: [00:01:00] Yeah, so I'm Johan van Amersfoort, Chief Evangelist at ITQ have been working with any infrastructure and cloud native infrastructure for the last decade or so. And have been working with several customers on AI engagements since 2018. Yeah.

But as part of a, an infrastructure engagement or of a cloud native engagement and not really specifically focused on AI as a practice necessarily. Yeah. In the last couple of years, we, we build out that practice and are now launching a, a label that's tied to the practice.

Ashish Rajan: And considering you guys have been in this AI space for some time. How, just to set some scene for people, a lot of AI projects seems to be chatbot, and I don't know how true that is.

So are you seeing more use cases outside of chatbots?

Johan van Amersfoort: To be fairly honest, I think the majority of use cases that we build for customers are not tied to chatbot. Okay.

Of course, since the rise of conversational AI solutions like ChatGPT and

Ashish Rajan: Anthropic and others. Yeah.

Johan van Amersfoort: Yeah. It's of course there's a rise there, but before [00:02:00] that we did a lot of traditional ML fraud detection platforms. Yeah. Computer vision platforms. Yeah. So object recognition in real time video, stuff like that. Oh, wow. Medical imaging. So there's a ton of AI related use cases beyond just like the conversational aspect of it.

Ashish Rajan: Yeah. Okay.

Johan van Amersfoort: And, all of them have different requirements, different aspects that make them somewhat complex. Yep. Or, have a different need for infrastructure or for platforms that run that can run them. Yeah.

Ashish Rajan: And do you find that building AI capable, because obviously a lot of people already have applications, they wanna make them AI capable.

And GenAI being the, to your point about ChatGPT, Anthropic, it's like it becomes this wave that's catching fire a lot faster than cloud did apparently. Yeah. Are people attaching AI capabilities to existing applications, or are we talking about newer applications?

Johan van Amersfoort: Could be anything. Okay. Interestingly enough, we see a lot of traditional ISVs adopt AI into their solutions.

Yep. But we also see, like a [00:03:00] couple of weeks ago, I was trying to figure out how fast I could build an application that does a specific thing. Okay. Make it cloud native. Yeah. And I use a coding assistant to to create, for instance we just had a, before the podcast recording, we had a conversation about CrossFit.

Yeah. I just wanted to find out how easy it is, first of all to create a kind of a training planner, training plan assistant, yeah. But also just dynamically adjusted training based on my skillset. So looking at one of the movements is a dead lift. Yeah. What if my dead lift is 280 pounds?

Yeah. And I want to improve on certain aspects of it, so I measure other weight. I could lift like a Romanian dead lift has a certain weight. And based on those numbers, I can adjust my training to increase or to get it more stable or whatsoever. And AI is a perfect solution for that.

Literally within a day I had it ready. Oh, wow. And just to use a standardized model. Yeah. Use a coding [00:04:00] assistant to build the application for me on a kind of per function basis. Yeah. And now it's it's ready. The best thing of all, it runs on a NVIDIA Jetson device. So it's a tiny device like A Raspberry pie.

Yeah. Okay. Basically on steroids with a GPU.

Ashish Rajan: Oh, wow. And to your point because the entry bar is quite low for anyone to do anything. Has there anything specifically that's come out of GenAI that's made it even more, popular. 'cause to your point, obviously now we have a lot more GPUs available easily, which was not the case earlier.

We have AI applications that are being built. Surely there's some challenges.

Yeah. There's a ton of challenges and looking at, so our headquarters is in Amsterdam. Yeah. We have a ton of customers in our geo and, I think one of the biggest challenges they're facing is mostly in the heavy regulated space.

So governmental agencies semi-government, public sector healthcare data must be stored within a secured premises. Yep. And nowadays we see a lot of what we call shadow AI, [00:05:00] where someone just has a credit card, consumes a service, puts corporate data in it, and it, it makes their work a lot easier.

Yeah. But it puts an organization at risk, so that's one. Yeah. Another one that we're seeing is that cloud costs because of that are significantly rising. Yeah. So being in control of costs, so being in control over where the data resides, but also ensuring GDPR compliancy is one of the, yeah, the challenges that we're seeing at customers and that can be simply solved to be fairly honest.

Oh, okay. And what are some of the solutions that come to mind as well? To your point, what could we, if they are starting a project today Yeah. What could they be doing to start on the right foot.

Johan van Amersfoort: So first of all a lot of companies are not considering private AI is what we call it. Yeah. As as a solution that could work. Yeah. Okay. Looking at the show floor here at KubeCon there's a ton of companies focusing on, I. Some form or some aspect of a private AI platform. Yeah. So looking at the different [00:06:00] forms of Kubernetes.

Ashish Rajan: Yeah.

Johan van Amersfoort: VMware is here, Red Hat is here, SUSE is here. They all have their kind of solution to run AI in the private cloud. Yeah. And it's still one of the things that companies are not seriously considering. When you think about AI, they immediately think about AWS, Microsoft, Google.

Ashish Rajan: That's right. Yeah.

Johan van Amersfoort: But without the kind of the, the certainty that, and the guarantee that the data they put in one of those clouds also stays within their geo or stays secure. You don't get that certainty.

Ashish Rajan: Actually, that's true because in the European context, especially now with the political thing going on everywhere as well, it's a lot more the data sovereignty is a big question as well.

Johan van Amersfoort: Yeah. And another issue that we seeing is that. The data you need to either train a model or to to get your AI solutions to work is increasingly growing as well. Okay, so look at, so one of the use cases that we built [00:07:00] for customers around medical imaging, if you take into account that a single microscopic image of a white field microscope could easily be half a terabyte in size, and you could create probably, I don't know, four or five images in a couple of minutes. So just thinking about having a couple of terabytes of images of a single patient, of a single occurrence of what they have and uploading it into cloud that's a massive issue. So we think that it's better to, to bring AI to your data instead of bringing data to your AI.

Actually, that's a good point. And so to your point, there are versions of localized private AI as well as people can use Yeah. And flourish with. Cool. That's most of the questions I had.

I've got three fun questions for you as well, that, ah first one being what do you spend most of your time on when not trying to solve the AI, private AI over the world?

You mean outside of work or? Yeah. Could be outside of work. Yeah. Oh, dude I spend at least 10 hours a week in a gym trying to stay healthy.

Yeah. Yeah. Fair. So the idea [00:08:00] is not to I know, become a bodybuilder or, so my kind of goal in life is to stay away from a, like an elderly home. Yeah. As long as possible.

Ashish Rajan: Yeah. That's a, that's a good goal to have.

Johan van Amersfoort: Yes. And spend a lot of time with my family. So conference season just started and last couple of weeks and we'll probably end the end of May.

Yeah. And which means that I'm away from home a lot. But yeah, as soon as conference season part one, like Q2 is done, then I'll be home for for the holiday season for a couple of months and I'll probably start traveling again since starting with the the end of August, oh, fair. Okay. Yeah.

Spend a lot of time with the family is fun as well.

Ashish Rajan: Yeah, fair. That's a good one. Which kind of leads me to the second one, which is what's your proudest moment that is not on social media?

Johan van Amersfoort: Ooh. Oh man, that's a tough one. I have several ones. I'm extremely proud of my kids, of my wife, who, you know, who manages the family while I'm away.

I wrote a couple of books and the first book that got released I did a book release [00:09:00] at VM World in 2019. NVIDIA bought a ton of books. They sponsored part of the book and part of the research. And there was handing over the book at their booth. Yeah. There was a line of, I don't know, 250 people.

Oh, wow. Just waiting to get a book signed. Oh my God. Taking pictures. That's probably because all of the, like a ton of work that goes into that. Finally seeing how the audience responded to that. That was definitely one of the work-wise, one of the proudest moments of my life. Wow.

Ashish Rajan: That's a good one to have as well.

Probably should ask you about your books later on as well. Last question what's your favorite cuisine or restaurant that you can share with us? Ooh.

Johan van Amersfoort: Oh man, that's also a tough one because I love a lot of food, different food. I love Current Favorite. Current favorite is probably, oh, that's a tough one.

Being Yeah. On the spot. Figuring out one. Aan de Poel that's that's it's a Michelin Star restaurant. Okay. They're in in Amsterdam. Yep. They're [00:10:00] really good. Good atmosphere like fusion food. Oh, really

Ashish Rajan: good. I will take the name for the next time in Amsterdam, but dude and where can we find you on the internet to connect with you and talk more about

Johan van Amersfoort: I run a blog. Okay. V-H-O-J-A-N dot nl. Okay. You can find me on LinkedIn. Johan van Amersfoort. That's probably where I'm most lively on still on X-V-H-O-J-A-N as well.

Yeah. Yeah, I think that's

Ashish Rajan: I will leave that in the show notes as well but dude, thank you so much for taking time. Yeah, thank you too. Thank you everyone for watching as well. Next time. Thank you so much for listening and watching this episode of Cloud Security Podcast. If you've been enjoying content like this, you can find more episodes like these on www.cloudsecuritypodcast.tv, we are also publishing these episodes on social media as well, so you can definitely find these episodes there. Oh, by the way, just in case there was interest in learning about AI cybersecurity, we also have a sister podcast called AI Cybersecurity Podcast, which may be of interest as well. I'll leave the links in description for you to check them out, and also for our weekly newsletter where we do an in-depth analysis of different topics within cloud security, [00:11:00] ranging from identity endpoint all the way up to what is a CNAPP or whatever, a new acronym that comes out tomorrow.

Thank you so much for supporting, listening and watching. I'll see you next time.

‍