Benjamin Tosado: Welcome back to Cleared for Takeoff. My guest today has a rare vantage point in our industry. He sees what’s actually being built on cloud infrastructure across aviation, manufacturing, and the defense industrial base simultaneously. Srinivas Matlapudi is a solutions architect for Amazon Web Services who covers some of the most technology-forward companies in those industries. We’ve worked a lot together on Trax and Airvoyant. He’s partnered with Deep Blue on those projects.

And he supports a growing number of defense industrial-based companies, right? Including StandardAero and General Dynamics. He’s been instrumental in guiding the projects with Deep Blue and AWS for Trax and for Airvoyant, including guiding the technology vision for the projects, helping us access funding and other types of support. So really glad to have you here, Srini. Welcome to Clear for Takeoff.

Srinivas: Hey, thanks Ben. Thanks for allowing me to be part of this podcast. Yeah, let’s go ahead. Thank you.

Benjamin Tosado: All right, sounds good. Yeah, we’re happy to have you here. So can you tell us a little bit about your background and how you ended up an AWS solutions architect and particularly covering this industry and these types of customers?

Srinivas: Yeah, sure. So my background is electrical engineering, My first job. I started as a sysadmin at Motorola as well. Everyone knows Motorola as like a cell phone maker, but at that time it had 150,000 people. It was deep, deep into defense, aerospace, government sectors. And it’s surprising some of my customers today were once divisions of Motorola that got sold off to the existing customers. So the question about aviation, manufacturing, defense, those roots go way back for me. So anyways, over there, I worked inside the data center focusing on core stack — compute, storage, network. So it was entirely before the cloud. So I get that vibe when I go and meet my customers who are today on premises.

Benjamin Tosado: Gotcha. Yeah.

Srinivas: I understand some of their thinking as well. I can put myself in their shoes, like, you know, their preferences that they want to still stay on premises. From there, my first cloud role was working as, again, infra as well as DevOps at Oracle Cloud. That was my first cloud role, managing about 500 virtual machines, both infrastructure as well as every service that went into the cloud. So the orchestration and all. And working over there, you know, a recruiter reached out to me from AWS. I just didn’t choose this path of getting into the aviation, manufacturing, defense space — it just happened that the group that hired me was already catering to those segments.

Benjamin Tosado: That’s it.

Srinivas: So once I joined, I really liked the challenges that they presented — the complexity, regulatory stakes, and all the real world consequences. I was kind of hooked because these are some of the hardest problems in cloud. And that’s what kept me here. Yeah.

Benjamin Tosado: Yeah, that’s awesome, Srini. So one of the cool things I like about producing this podcast is even with people I work with a lot, I learn things about them that I didn’t know, right? So I didn’t know you started in electrical engineering and then moved over to technology. That’s really cool. And I completely agree with you about the real world impact, right? I was writing a piece yesterday about how the geopolitical climate has changed significantly in the past several years.

There’s increased risk that’s introduced for our country, right? And if you look at that combined with the rapid acceleration of AI, we have more capability to protect ourselves, but our adversaries have more capability to attack us as well, right? So it’s a really important time for us to be doing this work.

Srinivas: Absolutely. That’s right, yeah, absolutely, yeah.

Benjamin Tosado: Yeah, yeah, yeah. So what is a day in the life look like for you, right? So what’s your day to day look like as a solutions architect covering these customers?

Srinivas: Yeah, this is an interesting job here. And AWS, it kind of puts you to the test as well. No two days look the same here. That’s what keeps it really interesting as well as challenging. In even a single day, I could be pivoting from one task to another — let’s say I could be in an architecture roadmap discussion with one customer, then maybe have to jump onto a project review meeting.

Or I could be showcasing a POC or a demo for another customer that I’ve built up. And then, you know, we spend a lot of time with our customers showing presentations, running workshops, because any new service that comes on board, we let our customers experience it. I know we’ve worked together with Deep Blue as well, with partners like you, showcasing those services, getting you on board. Training, providing the training. And between all these, I have to cater some time to upskill myself, because if I’m not up to date, I can’t help my customer. And sometimes my customers have critical escalations. I’m not directly in a support role myself, but I act as a bridge between our support team and the customer to escalate an issue or help them out.

Benjamin Tosado: Sure. Yeah. So it’s really like where the rubber meets the road from a technology perspective with the customer, right? So keeping them up to date on what’s going on, helping guide the vision for what they’re gonna do, make recommendations. If something happens and they need help, you’re not responsible for providing the support, but you’re bridging the gap and making sure they get to the people they need.

Srinivas: Yeah, absolutely. One point I would make — one of my previous managers, when I joined, I used to talk to him a lot. He would advise me, like, “hey, Srini, everyone has 24 hours.” So when he mentioned that, I kind of understood that in this role, you have to really prioritize. You have to be ruthless in prioritizing the tasks and only focus on those that have the maximum impact for the customers. So that’s a very important part for anybody looking to come into this role or join AWS — I would advise that.

Benjamin Tosado: Yeah, yeah, absolutely. Prioritizing and sticking with what’s important to the customer. Completely agree. So let’s talk a little bit about what you see in the industry, right? You’ve got manufacturing companies that make different types of defense and manufacturing equipment. You’ve got some aviation companies, you’ve got a software company that we work on together, right? So what do you see that they have in common when you look at the AI journey that these customers are going on?

Srinivas: I’ll speak in general — challenges are the same in most of these cases. Everyone is working on the same core stack. Everyone has challenges related to cost. Everyone wants cost optimizations. And then with respect to challenges, I see data fragmentation — for example, each of these industries, especially the manufacturing ones, have operational data logged in legacy systems — the maintenance records in one, supply chain in another, and engineering specs in another one.

Benjamin Tosado: Right.

Srinivas: Second is, I would say there is an ambition and reality gap between leadership and engineering. Sometimes it’s the reverse — engineering or leadership see the demos, they read the articles, they think it’s pretty fast to complete the task, but they’re not aware of the underlying challenges sometimes where the engineering sees that. And then they think, “how come my leadership thinks it’s easy?” And it could be reversed as well — when I go and meet the customers, the engineering group wants to move fast, but the leadership is not ready. And then the last one I would say is the regulatory permissions.

Benjamin Tosado: Yeah. Yeah.

Srinivas: Teams on the ground know what’s technically possible, but they’re also unsure what is allowed. So — is this compliant? Can we put this in the cloud? Will an AI-generated recommendation be accepted by a regulator? This kind of uncertainty creates those challenges. So yeah, I would see these patterns across the industry.

Benjamin Tosado: Right, right. Yeah, I would totally agree with that. So cost sensitivity, right? “Hey, how much is AI seems like it’s gonna be able to deliver these outcomes, but what is it gonna cost? And by the time we pay the cost, is it gonna make sense?” It’s absolutely some separation between expectation and reality across the organization, from leadership and from the folks on the ground implementing the technology. And in the DIB and in the aviation industries, there’s all kinds of regulatory boundaries that those customers have to deal with, right?

Srinivas: Yeah, I mean, when I go and meet these customers in person, internally with the groups and all, they cater their time, they show what their challenges are. You would be surprised — they want the solution. But they also have boundaries they respect. They have to respect what their security group is, teams internally, up the chain, looking at the restrictions that are placed for these guys. So you have to respect all those boundaries and come up with solutions sometimes that cater to everyone, both the leadership as well as the engineering team.

Benjamin Tosado: Yeah, yeah, absolutely. I know some of these customers in these industries are some of the most cloud averse, right? For years they’ve been looking at, “hey, we need to keep infrastructure on prem, we don’t want to leverage the cloud,” and a lot of that’s been because of the regulatory and security compliance requirements that they have. So have you seen that improve over the last couple of years?

Srinivas: It is slowly changing, and especially with the advent of AI — that’s a game changer. So it is slowly changing, but you still see that lethargy, that fear of change, and the fact that they want to protect their turf. I’m speaking freely here. And having that experience from the data center myself — being in that situation, when the cloud came in, I was myself part of that. So I understand their mindset and you have to approach it that way. You can’t just go in — sometimes I go with some of my team members and they’re like, “why, how come these guys are not changing? Cloud is the best, right? How come they have not jumped in the cloud yet?” But you have to understand that mindset — they have been like this for years, it’s a slow change. You have to advocate that slowly and introduce them to a hybrid. Bring one workload slowly, show them what’s possible, how much they can take advantage of — the agility, the scale, and everything.

Benjamin Tosado: So that leads me to another question. You’re working across several different customers and they’re at different levels of readiness to adopt AI and adopt the cloud, really. So what do you think is the difference between the most forward-leaning customers — the ones that are actually implementing cloud solutions, implementing AI solutions — and the ones that are kind of laggards? What causes companies to be held back? Is it regulatory, is it cultural — what do you think the differences are between the companies that are really taking advantage of this new tech and the ones that are being left behind?

Srinivas: So you’re asking specifically about artificial intelligence — the ones that are ready and not ready. So I would say the first one is foundational data. If the data is scattered across spreadsheets or tribal knowledge, still using legacy systems that have been untouched for a number of years, they are not ready. That’s a full stop. It doesn’t have to be perfect, but it at least has to be governed, connected, and queryable, because we have tools within AWS that can help them, and partners like you can jump in as well — Deep Blue can help them. That’s number one, I would say. And layer two is the cloud native infrastructure — you need elastic compute, modern APIs, and the ability to iterate quickly that cloud gives you. On-prem just doesn’t give you that agility anymore. Everyone that has been holding back so far — this is the inflection point. They can still try, they can spend six months procuring hardware, racking and stacking everything, but by the time everything’s deployed and they put a model in, you have the next model. These models just don’t have the capability to run on that hardware.

I’ll tell you an interesting example. Last year at the DC Summit, I walked into an Anthropic booth and I asked them, “hey, some of my customers are on-prem, they just don’t want to move — do you guys have any solutions for these guys?” They were just looking at me and they were laughing. No, like, “man, there’s no way these models can work at that level, on-prem doesn’t cut it.” And I think there’s a couple of others too, like organizational capability — are there people who can bridge that gap between data science and operations? So they could probably build a small demo that might work, but will never be adopted. And the last one I see is executive sponsorship, because AI definitely needs some runway. Let’s talk in aviation terms — if they’re expecting an ROI within like 90 days, it just doesn’t cut it. These guys need at least 12 to 18 months to prove the value, and they have to be protected from any revenue pressures. I see these patterns across for those that are ready and those that are not ready.

Benjamin Tosado: Yeah, that makes a lot of sense. We see some of the same things, right? We’re seeing AI as a driver for cloud migration now, for sure. I was talking to one of our partners at DC Summit this week — they really focus on our infrastructure migration business — and I was telling him we’re seeing more and more migration opportunities coming out of customers’ desire to leverage AI, and them understanding that running everything in a non-cloud, non-hyperscale data center makes it a lot harder. So I think that’s a hundred percent spot on. So what kind of mistakes do you see companies make when they’re trying to implement AI? Have you seen anybody have any false starts?

Srinivas: Yeah, I’ve seen a few, at least a couple of them. So building on top of broken data — we talked about how important that data is. I was recently at a boot camp and my colleague — every one of us here has to be specialized in one topic, I’m on container and also OpenAI — I was talking to another colleague of mine who was on analytics, and we were talking about how analytics was so hard before AI came along, right? But in all honesty, we both concluded analytics is still number one, because the data science and foundation for getting the right data to an algorithm, to produce the right model, which will generate the inference — without that, it’s garbage in, garbage out. So I’ve seen companies get excited about generative AI, LLMs, and agents and everything, but if their data is still in silos, ungoverned, you can end up with a small curated set that might run a demo, but it’s not ready for production.

And second is — we talked about this — companies that are fully on-prem are hitting that wall. The compute, the GPUs, the scale, you can’t rack that in their data center. So this is the inflection point — they have to embrace cloud, it’s not optional anymore. And third is confusing a POC with production. You can run it in a small, data-controlled environment, but you’ll realize that getting it into production is 10 times harder. So any of these can waste time — a year to 18 months — before they realize and have to think about going back to the drawing board.

Benjamin Tosado: Yeah, I think that makes sense. The whole data problem is something that comes up over and over again, right? And I agree — I think that’s why a lot of POCs don’t make it to production. You can build something cool that looks awesome, makes a great executive presentation on a really limited data set, but when you want to scale that out to an enterprise in production, you’ve really got to have the data platform situated, and you’ve got to be able to build a scalable, supportable enterprise product. A lot of these folks aren’t ready to do that. So we’ve got to do the right things first, right?

Srinivas: Absolutely. Yeah, I mean, I’ve also seen — just to rehash on that — some of the customers I talked to have not embraced AI yet. I still talk to some of them, they think this is hype, but they’re slowly getting on board, and they’re also faced with the challenges we just discussed. You have to spend some time curating the data and make it ready before you embrace it.

Benjamin Tosado: Yeah, yeah, absolutely. So let’s talk a little bit about our partnership with Deep Blue and AWS and how we kind of work together, go to market together. Can you talk about the co-funding benefit, right? Like, we were able to get, as an example, Trax and Airvoyant — the whole AAR subsidiary family there — a ton of funding. Can you explain how co-funding works at AWS and the different things you guys offer there?

Srinivas: Yeah, sure. So generally the co-funding, especially after the advent of AI, and when AWS initiated this — as we were all navigating through this landscape together, AWS introduced what we call a Generative AI Innovation Center, right? So this GenAIC team — their purpose is helping customers and partners like Deep Blue Cloud get started on the AI journey with real support. This is not just marketing — it’s actual funding, partner cash, hands-on engineering help, advisory support. Whatever is needed to go from proof of concept to production. But there are some caveats — there’s also a bar to meet, every project is not funded.

Benjamin Tosado: Right, right.

Srinivas: A customer’s project needs to generate at least about 500,000 in annual recurring revenue, right? So once that’s there, we put the proposal in — you give the technical approach, the projections, the feasibility — they review all of these. So it’s not just a rubber stamp, because they hold you accountable. Which is actually good, because it means the properly funded projects have a real path to production.

Benjamin Tosado: Yeah, totally. So I know that it’s all based on a return on investment, which makes sense, right — so you guys look at what this is gonna project in terms of AWS spend, and that kind of guides the funding. There’s some minimum thresholds, but you guys also do this for customers like Trax or Airvoyant, where you and the account team that supports them from AWS have worked with us and done a great job advocating to get them multiple projects funded at the same time as an example, which is not a typical thing that AWS does programmatically. So you guys believing in the product — or the project, sorry — and being an advocate makes a difference too.

Srinivas: Yeah, yeah, absolutely. I mean, it’s not typical for the same partner or customer to get multiple funding, and we are glad that Airvoyant got it, and Deep Blue was a big part of it. So thanks to you guys as well, on both Airvoyant as well as the maintenance planner, getting that funding. So the biggest message we want to give is: we are with you in this journey, and we’re not just providing infrastructure and models. I don’t know how other cloud providers are doing it, but we are completely on board saying, “we help you get started, be hands-on alongside you, and support you from that first POC all the way to production workloads generating real revenue.” That’s what co-funding means for us — AWS saying, “we believe in what you guys are building and we are investing alongside you.” So, yeah.

Benjamin Tosado: Yeah, yeah, absolutely. And that’s exactly how it’s worked at Trax and Airvoyant, and our partnership with you guys across those projects. I remember when you guys invited the stakeholders for those companies to DC, and did the first briefing with them, and then brought us in to sponsor a lunch and have a conversation with them.

Srinivas: Right, yes.

Benjamin Tosado: And that’s where we really got started. And now, fast forward a year, and look at all the progress we’ve made. Can you talk a little bit about how the three organizations have partnered together and how that journey’s happened?

Srinivas: Yeah, absolutely. So you’re talking about Airvoyant and Maintenance Planner itself, right? So I think Jon and Miguel came to us in January of ’25. Initially it was all — we talked about data so many times, right — initially it was getting all the data right. So building that data pipeline, ingestion architecture, and data flowing from their platform, the supplier networks, and on-prem customers into the curated format in the cloud — getting AI ready, that’s what we call it. And we were at that inflection point when we met in October, end of October of last year, for both these efforts — you guys were also there, so thanks for joining at that time, along with our leadership.

We brought in our senior most specialists, who delivered our key capabilities that solidified the message — I think Jon was still debating at that time whether to jump on AWS.

Benjamin Tosado: He was.

Srinivas: Yeah, I know — it was a big inflection point for us. We were trying to convince him, “hey, you guys are already on AWS, we have Deep Blue as partners as well.” And to get the co-funding, that’s when my manager advised me, during that meeting, to take these two efforts and talk to the GenAIC team, because until then I was not aware of that team. I tagged along with a colleague who had a head start, and we sped up the process, putting those reviews in, conducting those meetings together between you, us, and Trax and Airvoyant. And that’s where I think everything clicked, and I consider ourselves very lucky, and also the effort we put in together, the three of us, to get these two funded, and look at where we are today.

Benjamin Tosado: Yeah, yeah. We’ve got one product, Airvoyant. We helped them launch it in April, right? So that’s already a production system that they’re selling to airlines now. And over the next year, we’ve got our first couple of agents deployed there on Agent Core and Bedrock, and we’re developing six or seven more over the next year, and building AI foundry capabilities based on AWS’s AI-based development practices and tools.

Srinivas: Absolutely.

Benjamin Tosado: And that’s been an awesome story. And we’ve gotten started with Maintenance Planner Pro — we’ll be launching that later this year. So those stories are awesome. And we’re also helping them with a migration in their legacy business, and you guys have funded that with MAP funding — we’re helping them take their legacy customers and move them into the cloud.

Srinivas: Right, absolutely. And just to double click on what you said — going from just hearing about these agents to fully deploying those production-ready agents, sourcing thousands of suppliers and getting those codes — all thanks to you — getting there within a few months, and launching publicly with that brand name at MRO Americas this April, right? We were together, so I appreciate all your help helping our customers. We are partners in this, and we’re glad we have you on board helping our customer.

Benjamin Tosado: Yeah, thank you, Srini. And we’re really — I think you and your team are an example of exactly how AWS is supposed to work, right? You’re so supportive of the customer, so supportive of us and the overall partnership. And I’m really excited about those projects and the rest of the stuff we’re working on across your customer base, because I think we’re really working together to get the customer the outcomes they’re looking for, which is the way it should work, right?

Srinivas: Yeah.

Benjamin Tosado: That’s the way it should work. Yep. So that’s our big commercial aviation win, right? Let’s talk a little bit about the defense industrial base. So you’ve got GDLS in your portfolio that you support — sorry, I’m sorry, that’s my — man, I messed up the acronyms. You have GDMS in your portfolio, you’ve got StandardAero, right? So tell us a little bit about what makes cloud and AI challenging in the defense industrial base specifically — some of the regulatory challenges, that type of stuff.

Srinivas: Yeah, first I would consider not a challenge, but generally the way defense works — defense customers can’t always use the latest services, because commercial guys, when a service is released, can use it the next day. But defense has to, for any of these services, go through an ATO process — authority to operate. But having said that, AWS is still number one cloud service provider for DoD, now DoW. And we have the most number of services and features, if you look at services and scope, compared to any cloud service provider that has been ATO’d and is FedRAMP compliant. So we’re not just talking about GovCloud — defense has to usually go into GovCloud, we also have secret and top secret regions, for higher impact level workloads, whether it’s impact level four, five, or six — each with different service availability, compliance requirements, and operational constraints. And the third is the compliance stack itself — FedRAMP, ITAR, CMMC — these affect architectural decisions, like where data lives, who can access it, how it moves, and how it’s encrypted. And that’s where partners like you can come in and help the DIB customers, give those customized solutions that help them get to market faster. So I think that’s a little bit of a difference I see from the commercial aviation or manufacturing side to defense. But to be honest, defense, when it comes to AI, they’re pretty up there — they want to get the latest models, they want to try those out. There’s nothing holding them back except for the compliance.

Benjamin Tosado: Right. They want to adopt it. Yeah, that makes a lot of sense. And AWS — if you look at a company like General Dynamics or StandardAero, they’ve got commercial business, then they’ve got government business that’s impact level four and five, and then they’ve got impact level six stuff that’s secret and top secret. So you guys at AWS have your commercial cloud offering, which by the way has most of this compliance control coverage that you have in government cloud. Government cloud is just supported by US citizens only — it covers ITAR data and things like that. And then you’ve got the IL6 environment that’s completely separate. So you’ve got three distinct environments to help them across all the things that they need, right?

Srinivas: Yes. Top secret. Absolutely, yes. One point I would like to add — defense also presents you those customized challenges. For example — I don’t want to name this customer, they were my customer a couple of years ago — they present unique challenges. I was in a shop floor at one of these in-person meetings, they were looking for solutions, but they were also not ready to completely deploy in the cloud. They want a solution where cloud comes to them. We have a service called Outposts.

Benjamin Tosado: Right, yeah, like the off-prem, out-of-the-cloud environment.

Srinivas: Correct, it’s the on-premises rack that can be deployed, still managed by AWS. So you find some challenges — Outpost still goes into a data center, but these shop floors do not have a data center, and they were also against wiring — for some reason they said “you can’t just wire it.” So we also had to come up with customized solutions for modular data centers. There’s a commercial version of that also. But we always go to the customers and try to present them with any number of solutions — deploy these MDCs with a satellite link. And that’s where I think partners like you can also come in and help customers. So there are customized challenges you sometimes see on the defense side that you don’t see on the commercial side.

Benjamin Tosado: Yeah, yeah, absolutely. And I agree — AWS, you guys provide an awesome platform and all of this other support around it that we’ve been talking about today, and then we’re the last mile, right? We help the customer put it all together, stitch it together, customize it, and get it from that eighty to that hundred percent in partnership with you guys and the customer. So that’s exactly how it’s supposed to work.

And the defense industrial base environment is particularly interesting, because of all the compliance challenges and all the different things they have to do to leverage the same technology that we do. We have a customer right now that’s a startup, and they’re just trying to leverage Claude as an AI assistant for their users, and they can’t just go buy the commercial cloud that, as an example, we use at Deep Blue, but AWS has a couple of different options for them to be able to access the Anthropic models and use them, whether it’s going through Bedrock or using some of your managed services offerings.

Srinivas: Yes, yes, absolutely. Yeah. So I’d like to go back to that other question — we’ve seen defense adopt like I talked about, even the journey in the news in general. If you’ve heard — I’ve seen this where the Pentagon itself went from 80,000 people using commercial AI tools inside the Pentagon in December of 2025, to 1.5 million by June of 2026. So, and I’m seeing that in our other defense customers as well — they want the latest models, they want to get them on board as soon as possible. So there’s nothing stopping them, really — there are some challenges getting those models fast enough for these guys to use.

Benjamin Tosado: Wow.

Srinivas: And also, some of the challenges — the administration is also putting restrictions on which models can be used and which cannot. But AWS is also trying to bring on new models that can help them — like, we have a partnership with OpenAI.

Benjamin Tosado: Right. You were at a training for OpenAI, on the government environment, recently, right?

Srinivas: Yes, it was a boot camp that we had, getting our customers helping them on both those models if they have restrictions to use any of the cloud ones. So we’re enabling any model they can use — the latest ones — on our Bedrock platform, so they can get started and not be restricted.

Benjamin Tosado: Yeah, absolutely. We have a DIB customer recently that was telling us that was really important to them — being able to select a variety of different models using the same kind of platform, based on their needs, based on cost. You and I were talking about cost the other day — that’s always a factor in that as well, right?

Srinivas: Yes. Yeah, yeah, absolutely.

Benjamin Tosado: Yeah, cool, awesome. Well, we’ve got our rapid round here, Srini — we’re kind of wrapping up the podcast for today. So I’m gonna ask you a couple of quick questions if you can give me like a one or two sentence answer. What do you think is the most overhyped thing right now for AI in the industries and customers you’re supporting?

Srinivas: Sure. I would say — let me think about this — I think maybe fully autonomous agents in defense. Because if currently your process has, say, 10 tasks, fully autonomizing that into unsupervised decisions where consequences might be physical, I think that’s probably a little bit of hype. You might still need that human in the loop. It’s not a limitation, it’s a requirement. That’s probably what I would say.

Benjamin Tosado: Okay. So people’s expectations are a little bit higher than what the technology can bring to bear in that industry right now.

Srinivas: Right, I would say, yeah.

Benjamin Tosado: Okay. That’s awesome. What about the most underrated? What do you think people overlook the most?

Srinivas: This one is easy — data engineering and governance. I’ll go back to this again, because nobody wants to talk about it. It doesn’t make the headlines. It’s not the fun part at all, because everyone wants to just get the latest model and have the agents build everything. But I think it’s still the bottleneck for 80% of AI projects. And companies that invested in clean, governed data a year or two ago are deploying AI today. I think everyone else is stuck at step one, I would say.

Benjamin Tosado: It’s not the fun part — you just wanna turn it on and it’ll work, right? Yeah. Yeah, yeah, I agree, I think you’re right. And these are supposed to be rapid Q&As, but I think you’re absolutely right — customers that have been working on their data platforms for years are in much better shape right now. A hundred percent. I just had to double click on that one for a minute. So in terms of leaders at your customers or in the industries you support — what do you think they should be paying attention to right now on AWS? What should they be looking out for from an AI perspective?

Srinivas: I would say Bedrock and the managed AI stack — the barrier to deploying these models is dropping fast. Whether it’s guardrails or evaluations or Agent Core, it’s all becoming turnkey, especially with Agent Core — we recently also introduced a harness. And if you’re not experimenting with Bedrock today, I think you’ll be 12 months behind. And not just that — we also partnered most recently with OpenAI, getting those GPT models onto the platform. So AWS is making every model available, so customers aren’t short-changed, like they can only use a certain model with a certain cloud service provider. I would say that’s what they should be watching out for — Bedrock and the managed AI stack.

Benjamin Tosado: Advances in Bedrock and the managed AI stack. Okay, perfect. And last but not least — what does “Cleared for Takeoff” mean to you?

Srinivas: Clear for takeoff means, for me — everything: the data foundation is solid, the infrastructure is ready, compliance is handled, we have partners like Deep Blue Cloud in the picture, and the only thing left is to build something that matters. That’s all. That’s what clear for takeoff means for me.

Benjamin Tosado: Thank you. Awesome. Srini, thank you for joining us today. You’ve given a view of the landscape that a lot of people don’t get to see. I hope that what we talked about today makes some of the folks looking at deploying AI solutions think a little bit differently about what they need to do to get ready, and how they can partner with systems integrators and with AWS to go ahead and start their AI journey. For companies interested in building AI solutions and getting started with AWS, where do they go?

Srinivas: Always reach out to your account team first, and have a good relationship with the account team. Then we’ll introduce partners like you, Deep Blue Cloud, for actual implementation to get going. But start with the account team first, have a good relationship built up, even for partners and customers. That’s what I would say.

Benjamin Tosado: Awesome, Srini. So for our listeners, if today’s conversation sparked anything you’re interested in, come find us at Cleared For Takeoff dot fm. Until then, this has been an awesome Cleared for Takeoff. Thanks, Srini.

Srinivas: Thanks, Ben. Thanks for your time, it was great talking with you.

Benjamin Tosado: Yep, likewise. Bye bye.

Srinivas: Bye.