Innovation in DataOps and AI: BMC’s CTO on Orchestration and Data Efficiency | BMC Connect 2024

Hello everybody. I'm Mike Ard, and we're back at six five on the road and we're at the BMC Connect event in Las Vegas at the lovely Fountain Blue Hotel. And we're here with my old friend Rahm, and we're gonna talk about DataOps Rahm, welcome to the show.

Thanks, Mike. Always a pleasure to, uh, have a chat with you. There was this thing that happened a couple last year was this AI was everywhere.

It's a bright new shiny object. I think this year we kind of figured out it doesn't really work without data. Absolutely.

So, is DataOps kind of gonna be the new place where the cool kids are hanging out because the AI model isn't worth much without data? I, I think it's a fair statement. I mean, I'll go back to what I said in last year's connect events.

AI and data are in a cosmic dance where one without the other does not make any sense. One enables the other, they feed off of each other and jointly create value. So you cannot talk about one or the other in isolation.

So specifically a question about DataOps. DataOps is all about operationalizing your data management and data analytics use cases. It's great to use technology, but unless you can operationalize those use cases, you're not gonna be able to get value from them.

That's the premise behind DataOps. So it's a must have for ai. You and I have been talking about data management as long as I can remember What guilty as charged.

What is it about DataOps that's different than what we used to think of as data management in the first place? Great question. I mean, if you look at traditional data management, it's the collection of data, uh, ingestion, integration, storage, analytics, visualization, all those things.

What it doesn't account for is how do you get a handle from source to insights and keep everything in place such that if there is any break or flaw in the continuum from source to insights, how do you manage that? How do you mitigate the risk? And that's the key to operationalizing those use cases.

So when you deploy in a production environment, you want the rigor and discipline and enterprise scale resilience. And that's where Data Pipeline Orchestration comes in as an enabler of data ops. And with Control M, which is our, uh, industry leading solution in that space.

And Helix Control M, which is its SaaS counterpart, we have the best data pipeline, orchestration solutions, bar none, uh, for Enterprise Gate. Great customers. I also saw this new kind of data assurance tool that you guys are showing.

And what is that exactly? Yeah, So we, it had its genesis in the BNC innovation labs, which I established and I oversee. And as what we do in the labs is we solicit ideas from customers.

We, uh, look for their pain points and evolving needs, future strategy and direction. And we take a subset of those, synthesize them into ideas. And this was one that came from such conversations.

So data assurance, think of it as a compliment to data pipeline orchestration builds on the notion of business data observability. So think about applying observability concepts to analytics pipelines that through which data traverses. So if you can get a handle on the health and performance of data as it traverses complex pipelines and institute course corrections before the proverbial, uh, uh, things go hit the ceiling, then you're in a good place.

So that's the premise behind, uh, data assurance. Managing all that data is complex endeavor. Mm-Hmm.

Are we gonna get to the point where I'm gonna have something that feels like an AI agent to help me manage the data that I'm using to go build AI models? Uh, that's a great question, Mike. I mean, I think the short of it is AI and generative AI can be applied to pretty much every use case, but the trade off is what is the value?

What is the burden of implementation? Are you gonna get a return on investment? What are the associated risks?

Needless to say, uh, control M is ripe for, uh, Han ripe for augmentation with, uh, generative AI and ai. And that's definitely a key focus area for us, uh, to build on top of our existing automation capabilities with generative AI based, uh, workflows. Uh, we are working on another cool one, which, uh, exploits metadata using generative AI to provide previously unavailable insights to business services.

So those are many of the things that we are looking at with the power of AI and gen ai. Again, it's AI feeds data, data feeds the AI together, they are in a cosmic dance. And that's kind of how we're gonna operationalize all this.

Absolutely. 'cause one of the things you hear is that we don't have enough data engineers, but then the next logical question is, is well, do we need a data engineer for everything or can we democratize this to the point where mere mortals can do this? I absolutely think it can be democratized.

Uh, if we can institute, uh, code generation capabilities, uh, on traditional workflows for data specific workflows, you start with 50, 60% of the code or 70% of the code, you, your dependence on data engineers suddenly goes down. Don't get me wrong, we are never going to eliminate those people because the insights and the experience that they bring, the human touch is gonna be continue to be fundamentally important. But you can alleviate your own talent shortage, uh, uh, challenges with augmenting, uh, by augmenting the gen ai.

Do you think maybe the way it teams are organized is gonna change in the future? Because right now it takes a village to do anything. Yeah, and we don't have a village.

So, uh, I, I mean, uh, change is inevitable as the cliche goes, right? So it's absolutely gonna happen. In fact, it's already happening.

Uh, even within our own organization, uh, in our IT organization run by our CIO, they, they're re marshaling and pivoting resources and augmenting the gen AI so that the productivity gains are faster and the time to bring to market solutions stick, uh, solutions is that much fast, uh, that much shorter. So there are lots of folks out there and they're trying to figure out what their job is gonna be and how it's gonna evolve and what is their role gonna be. I don't think anybody's gonna be, be replaced per se, but what I was doing yesterday is not what I'm gonna be doing tomorrow.

Yeah. The, the nature of our work is going to change. The work itself is not gonna go away.

We are going to evolve and we are evolving towards doing higher value added work as opposed to a traditional, uh, types of work that we've become accustomed to over the last few years. So I think it, in some ways it is, uh, the techies version of an Iron Man coder. That's how I would look at it.

Are we gonna have a better understanding of what data is valuable? And I ask this question because a lot of times it people, they manage data, but it's often all the same to them. We now have data that is unstructured, semi-structured, structured.

Are we gonna be able to kinda ascertain which data has more value to drive into these AI models? That, that's a great question. So there's multiple, okay.

There's multiple dimensions to your question. Number one, you need to know that you're using the right data and the authorized data for each use case. Even something as simple as is this data, uh, authorized for public use in a or some other kind of, uh, broader use in training and AI model?

That's one set of considerations. But beyond that, specific to your question is it's not just about the business data. It's about the data about the data which is otherwise called the metadata.

This is data about code, about configuration, lineage, versioning, and so much more. This is pretty complex and dispersed across the technology layers, but you can use generative AI to link it to business services, understand interdependencies between your business, uh, data, your business services and data about the data, and that can provide killer insights and give you outcomes that were previously not possible. So absolutely do.

Where Is this going? All right, folks. You heard it here.

It's all about the data at the end of the day. And if you really want a job in ai, think about being in the data management side of the equation. You've been watching six five on the road here in Las Vegas.

We'll be back with some more episodes. By all means, check them out rom thanks for being on The show, Mike. Always a pleasure.

I'll leave you with this. Data is the fuel that powers ai. Alright, thank you.

Back in a minute.