THE LINUX FOUNDATION PROJECTS
Blog | I Am A Mainframer | Mainframe Connect

I am a Mainframer: Germanas Šamrickis | Teaching the Next Generation of Mainframers

By | November 17, 2026

Join host Steven Dickens in this inspiring episode of I am a Mainframer featuring Germanas Samrickis, a developer at Rocket Software in Vilnius, Lithuania, and a lecturer at Vilnius University.

Germanas shares his unique dual perspective on the mainframe: building mainframe monitoring products at Rocket Software while teaching the next generation of mainframers at one of Europe’s oldest universities. He discusses how students first react to z/OS, ISPF, and the 3270 interface—and how modern tooling such as Zowe Explorer, Zowe CLI, and VS Code is lowering the barrier to entry and making the platform more approachable.

The conversation also explores the value of “tribal knowledge” in an AI-driven world, why human experience and business context remain essential, and how observability and OpenTelemetry are shaping the future of mainframe monitoring. Germanas offers advice to his younger self—master the technology, but never lose sight of the business problems it solves—and shares his vision for a mainframe that becomes increasingly invisible while continuing to power critical business systems.

Watch full episode here:

OR

Transcript:

Steven: Welcome, you join us once again on another episode of the I am a Mainframer Podcast. I’m your host, as always—you’ve got to put up with me, I’m sorry—my name is Steven Dickens. And I’ve got Germanas from Rocket. Hey, Germanas, welcome to the show.

Germanas: Thanks for having me. It’s a pleasure. The fans of your podcast have watched lots of the videos, definitely. So yeah, we are very glad to be here today.

Steven: Long-time listener, first-time guest. I love it. So let’s get the listeners and viewers orientated. Tell us a little bit about your role, what you do, and we’ll jump off from there and go wherever we need to get.

Germanas: Yeah, sure. So I’m Germanas Šamrickis, a developer at Rocket Software in Vilnius, Lithuania, where I work on one of the mainframe monitoring products. So besides my daily job, I’m also a teacher—a lecturer—at Vilnius University, where I teach students everything that I know about mainframe technologies.

And actually, on that note, a fun fact that I want to share about Vilnius University: it was founded in 1579, which makes it one of the oldest universities in Europe. And despite being like 400 years old, it’s still teaching mainframes. So it’s like a funny combination, you know? Because sometimes people refer to the mainframe as being old technology, but the university is a lot older than that—older than the technology.

I think it perfectly shows that age isn’t what matters; what matters is whether something still delivers and continues to deliver value over time. So I think that, in my position, I see myself occupying two seats: being a developer and being a teacher and lecturer at the university. So I get to see both sides of the industry—the systems that power the business, all the ATM transactions, all the health insurance transactions, etc.—and the people who are going to run that technology, who will make decisions in the future about that technology and where it’s going to go next. So, yeah.

Steven: I mean, it’s those two hats that you’re wearing. Let’s maybe follow the teacher’s thread of the teacher for a moment. You’re, I’m assuming, exposing kids to the mainframe who’ve never seen the platform before. The first time they’ve seen a 3270 screen is potentially in your class. What’s that sort of “hello world” welcome kind of moment for those students?

Because I think a lot of people see barriers to younger professionals picking up this platform. I don’t see that. I think people are curious, and almost more curious about something that is a bit harder to get access to and a bit more unusual. If it’s kind of, “just jump on the web and it’s really easy,” some kids gravitate towards that. The kids like the “this is a bit harder to get into.”

Do you see that same experience? I’d love to—there’s probably not many people I get to talk to about the first time somebody logs onto a mainframe. So I’d love to understand that in a bit more detail.

Germanas: Yeah, well, it’s actually the very first question that everyone asks me about when they’re in a conversation or somewhere at some kind of conference. They have to ask me, when they realize that I’m also a lecturer at the university: “How do students react to the mainframe?”

And the honest answer is: probably most of the students are surprised about the technology. They definitely have a different vision and understanding before joining the classes. Maybe Google was the mainframe, and one of the first big pictures would be someone’s big, big rack somewhere, occupying a whole building or a whole room. So they have that vision. And some of them have heard—or maybe did some more digging into it—that it’s actually not the whole technology, but that it’s modern and up to date.

But, all the time, every class has its ups and downs. For instance, everyone struggles when they need to log into the system for the first time and use, for instance, a 3270 screen, go to ISPF navigation. This is something that is outside of their comfort zone, and they find it a bit strange. But at the same time, some of the students find it very nicely done, efficient, and very intuitive—where to go, how to navigate. So it depends on the person.

Usually, yes, shocked—but then, when they get familiar with the system, they appreciate it. And especially right now, with the newer stack—Zowe Explorer, Zowe CLI—that definitely lowers the barrier for entry to the mainframe. The students can usually get up to speed pretty quickly using VS Code and the CLI, because I think this is what they are comfortable with and have been using for a long time already.

So now I’m trying to do more and more of that modern stuff. I’m not going full “troubles and headaches with remembering navigation for 3270 screens,” but more focusing on how to get things done more easily using Visual Studio Code. And when we do that in the class, the quality and the speed—how fast we can go through all the topics in the class—definitely ramps up.

Steven: Zowe made it—I mean, Zowe is what it is, it’s 67 years old now. Does that make a material difference to the experience that kids are having first off with the platform?

Germanas: Oh, yeah, definitely. I would say that definitely made a difference. At the end of every semester, the students used to get that question about the subject itself. And for a few years—maybe four or five—I didn’t know exactly how the responses from the students had changed over time.

And taking into account that, during those years, I started to push more and more of the modern stack on the mainframe, I saw that the confidence in learning increased, and the satisfaction overall with the experience of using the system definitely increased. So that made a huge difference for the younger generation.

Steven: So one question we’ve been asking in the  last four episodes—sort of the fifth one—we’ve been asking about AI. You’re operating at 24 by 7, you can just ask it a question. What part of the mainframe can only be learned by a human? You’ve got a unique perspective there, I think. In a classroom, you’ve got students learning on their own, and they can go and ask AI questions. What can you only learn from a human?

Germanas: Well, I think there’s something called tribal knowledge.

Steven: Does that make you a tribal elder?

Germanas: Yeah, probably.

Steven: That’s a compliment.

Germanas: Yeah, thanks for that.

So, yeah, that tribal knowledge—and it’s not only in the mainframe area. I think that everywhere in the software industry, there is some kind of tribal knowledge that you cannot replace with AI. And it’s probably down to the fact that some of the most important decisions in companies were never written down.

So probably a senior engineer would see a piece of code and think, “I understand that it’s kind of inefficient, but it exists for a very specific reason—to back up 20, 30, or whatever years of business logic.” So AI can analyze the code, but it cannot know or understand the history—why it’s there. And that makes a huge difference.

Yes, as a companion, AI is a very good tool, I would say. It definitely could be there for you to understand and get up to speed with some basic principles, explain, for instance, obscure code, search through decades of documentation in a snap of a second, or give you some great examples to start with. So AI is definitely very powerful. But again, tribal knowledge is something that you cannot replace anytime soon.

Steven: So let’s pivot and move a little bit to some of the work you’re doing at Rocket. You’re a huge contributor to the Zowe project and the open community in general. So what do you say you’re a developer for Rocket—what are you working on, and how do you do it in the open? Is it on Rocket products? Is it a combination of the two?

Germanas: Well, I’m working on a mainframe monitoring product. So I’m developing the monitoring tool for the customers, and the development falls down into different categories—whether it’s something newer, new UIs, or more traditional, effective 3270-driven UIs. So it’s a different type of stack that we’re working with.

Steven: So where are you seeing that go? Is OpenTelemetry something that’s coming through in the project, and the monitoring product—a long-standing monitor? Do you see that kind of changing? You’ve been involved in telemetry, you’ve seen OpenTelemetry and observability coming to the overall space. Are you seeing it evolve, and people are not sort of looking at new paradigms?

Germanas: Well, yeah, definitely. From conversations with customers, they’re definitely looking into OpenTelemetry to get an end-to-end picture. Before I was working as a Rocketeer—how we call ourselves here—I worked as a client of Rocket, for one of the banks. So we had the problem back then of understanding our infrastructure: understanding where a transaction originates, how it goes from which components on the network to the mainframe, and what the loopback was.

We tried, back then, only in the bank, to build—how we call it—“map” the flow, to understand the cycle of the application. And, to be frank, I don’t think that we were successful in that. It was too much to handle—too many things that not everyone knew, how everything ties together.

And OpenTelemetry would definitely be something that could help clients map application logic and the flow through the network with different components. But again, it’s something that’s still being actively developed as a standard now, and we will definitely see things change.

Customers are looking into that, definitely. Some of them have already implemented some OpenTelemetry solutions for some of their applications; some of them are just looking cautiously at it, always considering the cost, because it definitely adds additional cost for your monitoring. So it depends on the clients.

Steven: So, we talked about it—you’ve been listening to the show. I always ask this question, these next questions, of guests. I love your perspective, because I think there’s not many people who are as connected back to younger professionals as you obviously are.

The question I always ask is: what would you go back and say to yourself, if you could give yourself some advice from you at 20–21, in college, kind of learning this platform? What would you do, and what advice would you give yourself? And you’ve probably given that advice to a bunch of students, but I’d love to hear it framed in what you’ve said to yourself. What would you go back and say to yourself?

Germanas: Oh, yeah, that’s the tough one. There is lots of…

Steven: You see the same question for ideas, it’s a great question.

Germanas: Yeah, yeah, yeah, definitely.

So I think that what I would tell myself is: master the technology, but never forget that it’s only a vehicle. The real destination is to understand the business problems and how to solve them.

I know probably lots of folks in their earlier years spend a lot of time and energy on understanding programming languages, platforms, and technical details. And that’s perfectly fine—those things definitely matter—but they change constantly. What creates long-term value is understanding why those systems and things exist.

Steven: Look a little older than just a few of them.

Germanas: Yeah, exactly. So you need, all the time, to try to look broader—to have a broader perspective and horizon—and understand what happens when it fails, what will be the impact back then.

Steven: I mean, we’re talking a lot about context and perspective at the moment, but I think what you’re saying—and what I’m hearing—is that the context around the technology is almost as important as the technology itself. Am I hearing you right?

Germanas: Yeah, yeah, yeah, definitely right about that, yes.

Steven: Okay. So I always ask this question as kind of my final question; it’s served me well over these years. Where do you see this platform five years from now? We’re going to get a new box, we’re going to get new operating systems. I’m asking you to kind of look a little beyond that and be able to give me a perspective through some of the work you’re doing on the platform, through some of the students that you’re talking to. How does that sort of form into an opinion about where we’re going? Where do you see the mainframe in 2035, maybe? Let’s look a little bit further out.

Germanas: Well, it’s going to be easy, you know, if I would have the crystal ball, I’m definitely going to be a millionaire.

Steven: You don’t get to, with the last question, you don’t get to go back inside.  you don’t get to stockpick going forward. You don’t get that. I’m sorry, this is a technology question.

Germanas: Yeah, You know, definitely the platform is not going anywhere. It’s not going to die in the near future—because it’s nothing that I’ve heard since I joined the industry. But now I actually hear those voices less and less, because of the newer developments on the mainframe side: quantum-safe encryption, AI on chip to monitor all of the fraudulent transactions on the chip.

And the latest news that I think came out last week or this week about dual architectural support from Marvell—that will definitely make the platform even more powerful. And I think that it will be… because now you don’t hear about the mainframe unless something very bad has happened: “the banking system stopped,” and all this is a thing back from the mainframe.

I think that the platform will still be there, but it will be even more increasingly invisible, while serving very important business purposes. That’s definitely true.

Steven: I like that: invisible and important. I think that’s a great way to wrap us up here.

Well, you’ve been watching another episode of the I am a Mainframer Podcast. Please click like, subscribe, share with a friend, do all those things. The more you subscribe, the more that this helps promote this podcast. So please do that. We’ll see you next time. Thank you very much for watching.

Germanas: Yeah, thanks.

Narration:

Thank you for tuning in to the Mainframe Connect Podcast and this episode in the I Am a Mainframer series, sponsored by Phoenix Software International. Like what you heard? Subscribe to get every episode or watch us online at openmainframeproject.org. Until next time, this is the Mainframe Connect Podcast.

Share