AWS Summit DC 2026 | Francessca Vasquez, AWS

Clips
More from AWS Summit DC 2026

Francessca Vasquez

VP of Professional Services and Agentic AI

AWS

play_circle_outline AWS $1B Forward Deploy Engineering: Accelerating Enterprise AI with Measurable Outcomes and Teach-to-Fish Upskilling

play_circle_outline Agentic-First FDE: AI-Native Development for Compact 5–6-Person Teams with Embedded Customer-Site Agents

play_circle_outline Hiring profile: multifaceted engineers—curiosity, problem-solving, AI-native development skills

play_circle_outline Common Methodology and Reusable Patterns for Scaling Agentic Systems with Semantic Layers, Knowledge Graphs, and Data Foundations

Info
Transcript

Francessca Vasquez, AWS

Francessca Vasquez

VP of Professional Services and Agentic AI AWS

In this interview from AWS Summit Washington, D.C. 2026, Francessca Vasquez, vice president of frontier AI engineering and services at Amazon Web Services, joins theCUBE + NYSE Wired's Gemma Allen to discuss how AWS's newly launched Forward Deploy Engineering initiative is redefining what it means to take enterprise AI from experimentation to production. Vasquez traces the path from AWS's Machine Learning Solutions Lab in 2017 through its Generative AI Innovation Center to the current FDE model, built for the agentic era. She explains how embedded agile pods ... Read more

explore Keep Exploring

What does AWS's new $1 billion Forward Deploy Engineering business unit do, how does it differ from traditional consulting, and what will a typical engagement look like in practice (including timelines and goals for customer self-sufficiency)? add

What types of people, roles, and skill sets does AWS place into its Forward Deploy Engineering (FDE) group, and how are those teams organized? add

What skills, qualities, and mindset are most important for an FDE working with AWS customers, especially on AI-first projects? add

How is AWS organizing its teams and development approach to help customers accelerate adoption of frontier AI by leveraging shared methodologies and customer context (semantic layers, knowledge graphs, taxonomies/ontologies)? add

bolt Powered by CUBE AI

Francessca Vasquez, AWS

search

Gemma Allen

>> Welcome back to theCUBE. We are live here at AWS Summit in Washington DC 2026 to bring you all of the action talking about this new paradigm for AWS and what this agentic world will look like. Join me now for a conversation on exactly that. Hot off the press with some very exciting news is Francessca Vasquez, VP of Frontier AI Engineering and Services at AWS. Welcome, Francessca.

Francessca Vasquez

>> Thank you so much, Gemma, and thank you to theCUBE. I'm so excited to be here today.

Gemma Allen

>> We are delighted to have you, especially on a day like this because you are part of an organization that just got some great stage time on the keynote with some big announcements. Maybe let's start there. Talk to me about what you guys announced today and what it means for you and your org.

Francessca Vasquez

>> Yeah, it's such a great question. So today here at the Washington DC AWS Summit, we were able to announce that AWS is investing one billion in launching our forward deploy engineering business unit. And why that matters is because we have so many enterprises that are seeking our help to help them operationalize AI and get a lot of value and do so with compressed timelines and speed. And so I'm very excited that we're launching here with governments, with private sector organizations. It's going to be a great journey for those customers. We're thrilled.

Gemma Allen

>> If we heard a $1 billion investment in human capital and consultancy five years ago, I would have thought that's a lot of PowerPoint spend, right?

Francessca Vasquez

>> Yes.

Gemma Allen

>> We know that that is no longer the case. The challenge that you guys are looking to solve is, it's an evolving challenge and it's so unique to any technical challenge we've ever seen before. We're in this space where it's all about speed and impact and outcome. And there has been a lot of obstacles for many organizations to making that happen and to meeting those market pressures for such a long time. Talk to me about what's different about this org. Like what does this consultancy engagement look like, feel like, run like? Bring it to life for us.

Francessca Vasquez

>> Yeah, it's important. So by the way, the one billion investment, it's a big investment.

Gemma Allen

>> Huge.

Francessca Vasquez

>> That investment feeds people that we currently are bringing together from internal groups within AWS. It is for market development funding and it's also for our very rich partner ecosystem where we intend to have a lot of enablement for the partners that will be working with us on FDE. For our customers, we've been thinking about this for a long time and as you can imagine, when we started building capabilities for companies that were just getting started with AI circa 2017, we ended up creating this machine learning solutions lab and it was like very foundational to get customers started. Three years ago, those same customers said, "Generative AI has hit and now we need help with use cases and getting those use cases into experimentation and experimentation into production." And so we created a GenAI innovation center. And now we're at this incredible inflection point where agentic systems or where customers are really trying to embed AI into their workflows to drive business value and they want for deployed capability. They want embedded engineers who understand how to build with speed, who are looking to help them get to production in a matter of days and weeks and not months and years. And the most important thing, we don't view this as traditional consulting. We view this as deployment, outcomes, highly compressed timelines and for customers we hope that at the end of our very fast 45-day sprint, they can become self-sufficient. And that is for them, that is upskilling their talent. So we're very excited.

Gemma Allen

>> Wow. Teach-them-how-to-fish model.

Francessca Vasquez

>> Yes.

Gemma Allen

>> Not to use the Bible reference at a tech show, but it's the best reference I can come up with.

Francessca Vasquez

>> Yes. And have them co-build so that there is active skills translation happening.

Gemma Allen

>> So help me understand the profile of some of these customers, right? Because I think we're at an interesting time in tech where suddenly everyone's expected to be a visionary, everyone expect to be a technologist, one step ahead.

Francessca Vasquez

>> Yes.

Gemma Allen

>> We know that there has been challenges in the tech landscape for a long time, challenges around structured data, challenges around knowledge share, knowledge that lives in people's minds, not necessarily documenting. All of these problems, they didn't disappear overnight as we began to talk about AI and agentic. Talk me through what a typical customer would be for this and how they're solving, I guess, too, Francessca, for some of those challenges that have kind of haunted industries and segments for as long as they have.

Francessca Vasquez

>> Let me take the first part of the question, which is like the type of people that we at AWS are putting into our forward deploy engineering group. I think the first thing is we are highly indexed on sort of product outcomes and really making sure that we're delivering value fast. And we believe that the how we help these customers change the way in which they're working is through taking traditional methods like software development lifecycle and making that AI-native. And so AWS has actually established an AI-native development lifecycle capability from how you plan and ship software, how you deploy, how you resolve, all AI-natively. So our team of FDEs, they build natively, AI-natively, they also build with agents and we do so by leveling up our software development engineers. We in some cases of our use cases have applied scientists and then we put these teams into agile pods and they deploy at the customer as a unit. And so the team is normally five to six humans, human engineers plus their AI agents and that sort of gives them the how. I think for the customers, there's not just one industry that gets value out of FDEs. I think where we see the best patterns are those industries that have very strong leadership conviction. So they've got a C-suite or a board that's very convicted on wanting to achieve upended outcomes for their business workflows. We see organizations that have prioritized data. Turns out to do AI really, really well, you really have to have a good data foundation and the industries that have had to deal with data for long periods of time tend to be industries like financial services, governments and state and local institutions, healthcare as a discipline and even media, entertainment, games and sports. They just tend to be data rich. So we find them to be very good candidates for FDE value. And then the last thing that I'll just say is when you have very good FDE deployment and we focused on the last mile, truly getting to production, it doesn't mean that you don't have to pay attention to the things that customers need the most. And those things are like, how do they think about their operating model in the age of AI? What does security look like in the age of agentic systems? What about governance? And so all of those things still matter. The FDE's benefit is that they're building in the customer's environment with the customer's constraints. And so it's really about meeting our customers where they are and not just building in a lab or isolated and then handing over some package to a customer. We really get involved in making sure that we are embedded in their workflows.

Gemma Allen

>> And what do the existing environments look like when we talk about embedding those workflows? Are we talking Bedrock, Agency, Amazon Q? What sorts of pre-reqs or what parts of this stock typically tend to be in place already to optimize for agentic?

Francessca Vasquez

>> Well, the customer patterns that we've seen are pretty much our customers that are building on Bedrock and we have a huge advantage in that we're model agnostic. We meet customers where they are. So sometimes they like to deploy first-party model with AWS on Bedrock. Sometimes we work with our frontier AI lab partners like OpenAI and Anthropic building on Bedrock, that works equally too. Some customers may be at a point where they want to start with AgentCore or they're starting to do spect-driven development with Kiro. Doesn't matter. We've built a lot of different extensions for the FDE organization to be able to leverage when they're building with customers, which makes the choice for customers much more valuable because they feel like they're not locked in, they have options and then we can build to any of the workflows. In healthcare, it could be completely related to like claims processing and financial services. It could be like a fraud loss prevention workflow. That's all unique to the customer and their industry.

Gemma Allen

>> I think that's very important when we think about vendor lock-in as a challenge for the last three decades of tech, right?

Francessca Vasquez

>> Yes.

Gemma Allen

>> Whatever paradigm comes next has to have a level of agility or who knows what sorts of challenges people could face. I'm interested in this skillset of the team because when we think about an AI engineer, somebody told me recently it's like claiming you're a theologist, right? Like what are you actually an expert on? What religion speaks to you? And it is an interesting time. I'm sure it's a very varied skillset, but are these like data folks? Are they machine learning experts? What's the typical profile even you're hiring for?

Francessca Vasquez

>> They are so multifaceted, but we feel very fortunate because AWS has created a host of builders who have very deep curiosity. They have very deep problem solving. There's a mix of both technical and soft problem solving skills that are very necessary for an FDE type role. I think every customer is different. So sometimes in our five to six person pod, we will be very good at being able to ship fast capability and sometimes we'll couple that with the customer who wants very deep industry, like people that have ship and deploy capabilities within safe financial services as an example. So I think the most important things are you want FDEs that are deeply curious. A lot of the problem solving is not determined, it's not defined before you go work with the customer. Sometimes you are truly discovering and deploying all in one sprint with this customer. And so that curiosity, the ability to be creative, to problem solve, and we want people building AI-natively so you have to be willing to actually reimagine how you deploy with an AI-first, with an agentic-first mindset. And we provide tools for that, but it's FDE gets the customers changing their ways of working during the deployment and so they learn very fast.

Gemma Allen

>> So what does a typical engagement look like? You mentioned like almost a microcosm of an agentic workload, people with different skill sets working inside a customer or on a customer site.

Francessca Vasquez

>> That's right.

Gemma Allen

>> But clearly side by side with a lot of folks because a lot of this needs context to run, right?

Francessca Vasquez

>> Yes.

Gemma Allen

>> And you mentioned I think 30 days or 45-

Francessca Vasquez

>> 45 days.

Gemma Allen

>> 45 days. That's impressive.

Francessca Vasquez

>> That's the framework that we came up with.

Gemma Allen

>> I mean, we're all used to the world of consultancy where somebody presents you a PowerPoint of a deck in 45 days with some great ideas, right? But to promise outcome, it's impressive, it's ambitious. What does it look like?

Francessca Vasquez

>> Yeah. So we've found that in working with customers, they had had this experience with sort of traditional consulting thinking where they often thought that their outcomes and transformations could only happen in extended years. And I don't think any organization today can afford to drive value in years. Things are moving so fast, the pace of innovation, they need it fast. And so we came up with this very simple construct to say, speed is the new currency for a lot of customers. How can we get customers thinking in faster sprints? And so my team, they came up with this model of we'd really like to get ideas out on the table in 45 minutes with the customer. After that, can we get to like validation of what we can actually go drive in a proof point in 45 hours? And then after that 45 hours with the customer, how do we then go develop and decide we're going to ship something in production and value in 45 days? And what that does is it's very ambitious for most customers. If we do it faster than 45 days, great. We all get a gold star, but it really gets them thinking about what they can actually deploy and implement in a possibility stating instead of thinking that they can't derive any value over a longer period of time. And so for us, we like pushing on that sprint and then for them, again, a great outcome for us is if the customer is also becoming self-sufficient in the process, which is one of our stated goals. So compressed timelines, them getting self-sufficient, live in 45.

Gemma Allen

>> Wow, I love it. That's a great marketing tag.

Francessca Vasquez

>> Yeah.

Gemma Allen

>> Live in 45.

Francessca Vasquez

>> New deployment. Live in 45.

Gemma Allen

>> I mean, it's catchy. It'd make a great jingle.

Francessca Vasquez

>> Yeah.

Gemma Allen

>> So great news today, exciting news. Pressure's on now, right, to bring this to light. Like you come off that stage and you think, "Okay, a year from now, what are we going to be talking about?" What does that look like? I mean, what are the priorities now? Are you scaling? Are you hiring? What does the next six months look like for you, Francessca?

Francessca Vasquez

>> Yeah. So couple of things. I think one, we have now curated all of the various teams within AWS that were contributing to some element of FDE. So now they're in on single business unit that will allow for common context, repeatability and reuse patterns that allow us to go very fast. The second is we have a common methodology. We use the same AI DLC development for our internal building of AWS Service. Now we're bringing that to customers. And so customers are getting the best of our ability to do frontier AI engineering directly with them. I think third, you mentioned this earlier, context matters and so many of these organizations, they will have so much more new context, semantic layers, knowledge graph. Pick your taxonomy, ontologies. We want to put all of that great IP engineering that customers will get in their hands. And so we're very excited to accelerate their ability to leverage the power of AI through data to have a ton of context composability in the engagements that we'll support. And then the last thing that I would just say is, now that the team has been sort of formalized, we expect to see the demand. We want to get as many customers in the funnel to support them as fast as we can so that in the next year, Gemma, when we're back at the DC Summit of 2027, all of the customers that you'll get a chance to hear from will tell you that they now actively are operating like frontier AI companies. They've seen value now in 45-day sprints versus 18-month or 24-month projects. They have upskilled their talent in the process. They have an ecosystem of AWS trusted advisors and partners that they've used to help them get there. And then lastly, we've delivered on the outcome that we all mutually said we would do on their workflow.

Gemma Allen

>> I certainly look forward to-

Francessca Vasquez

>> That's success.

Gemma Allen

>> Listen, I'm backing your bet all the way. I look forward to hearing from some of those customers. And you know, you touched on some important things there, hiring brilliantly minded people, keeping humans in the loop in all of this and then teaching someone to fish, right? That is true partnership. So Francessca Vaquez, thank you so much for joining us on theCUBE.

Francessca Vasquez

>> Thank you for having me. Great to see you.

Gemma Allen

>> I'm Gemma Allen. We are live on the queue at the AWS Summit in Washington DC 2026. It's an exciting week here in DC and we have some fantastic conversations coming up in the show. Don't go away.