In Conversation with Raúl Nogales, Founder and CEO, AVAP

We recently had an informative conversation with Raúl Nogales, Founder and CEO of AVAP, where we discussed his lifelong passion for programming and the corporate and entrepreneurial journey that ultimately led to the founding of AVAP and the development and introduction of AI Systems – a new software architecture that is redefining how AI software is built by human–AI teams. It arrives at a moment when, in Nogales’s own words, most AI solutions on the market today are little more than wrappers around ChatGPT or Claude.

1.     Your technology leadership journey is truly inspiring, spanning multiple countries and the founding and leadership of several technology companies. Could you walk us through this extraordinary journey that ultimately led you to found and lead AVAP, a company transforming how AI systems are built and operated in the AI era?

I could tell you a great deal about it. It would take many pages, many hours of conversation, and more than a few cups of hot coffee to tell the story you are asking me to tell. But, as my Mexican brothers would say, let’s make a long story short. And if you really want the long version, you can always read my latest book, A Orillas del Aqueronte, where I tell this entire story in detail.

But let’s start with the basics: technology has been my passion since I was a child.

I was lucky enough to witness the birth of home computing, and even luckier to have the curiosity and talent to turn that passion into something tangible. At ten years old, I was already programming in BASIC, experimenting with assembly language, and making the most of a friend’s Amstrad CPC464 with a green phosphor screen. At the same time, I was eagerly waiting for my first computer at home. When it finally arrived, it was a 32KB Sony MSX 101P of 32KB.

My hyperactivity has always pushed me to discover new worlds and constantly take on new challenges. Some came out of necessity; others, I simply invented for myself because I needed a new challenge. But I can say that I have always had an entrepreneurial spirit, even during the early stages of my corporate career.

Today, in many ways, that story is repeating itself. The passion never went away. It continues to grow as I watch technology, this time under the banner of AI, change the world. And I feel incredibly fortunate, and deeply privileged, to be part of that transformation.

Back in the early 2000s, after more than a decade already working professionally, I decided to leave the corporate world where I had built my career, primarily in financial technology.

By then, I had already had the opportunity to work on and participate in international payment projects such as MobyPay in Spain and mPesa in Kenya. Eventually, I decided to become a full-time entrepreneur.

One of the last, and most important, corporate projects I was involved in was Hal Cash. It was not an entrepreneurial venture in the traditional sense, but it gave me a unique opportunity in 2008, during a very difficult year in Spain.

The Spanish banking partners, ING and Caixa Galicia (now Abanca), among others, approached me with a proposal to launch Hal Cash Mexico as a 50/50 venture.

Looking back, I can say that deciding to leave Spain in 2008, in the middle of a major banking crisis, and launch Hal Cash Mexico was probably the decision that shaped and determined my future, and, to some extent, the future of the companies I would go on to build over the following twenty years.

Hal Cash offered a fantastic service: you could send money from a bank account to an ATM and withdraw it using your mobile phone, without needing to be a bank customer or even have a card.

But it was ultimately a product that worked extremely well in Spain because of a number of very local conditions. All of the partners were banks. The commercial network was made up of bank branches. And, most importantly, the product existed within a society and a country that was essentially fully banked.

The idea of bringing Hal Cash to Mexico was not a bad one. But, looking back, I would say that our approach was not quite right.

I found myself in a country with a very low level of banking penetration, where it was simply impossible for money to flow from a bank account because there were not enough bank accounts in the first place, regardless of what government institutions, politicians, or official figures might have claimed.

A year later, we had secured commitments for five million dollars in investment from different local and international organizations. But there was a clear lack of continuity in the support from the Spanish banks, largely because of the crisis they were going through.

So, I decided to sell my 50% stake to an American investor based in Miami who was originally from Mexico and understood the country’s idiosyncrasies extremely well.

I took the proceeds from that sale and decided to bet on a new project, one that would allow me to turn my vision of what Hal Cash could and should become into reality.

That is how Pademobile was born.

It was born out of the realization that I could not take Hal Cash where I believed it needed to go. I wanted to move beyond being just another link in the last mile of local and international money transfers and remittances, where there were so many structural limitations, and instead build the first digital banking platform primarily focused on financial inclusion.

It is important to understand the scale of the opportunity. The U.S.–Mexico international remittance corridor is the largest in the world, with approximately $60 billion in money transfers every year.

Between 2009 and 2015, we grew to five million registered users. We developed financial inclusion initiatives such as government benefit payments in Campeche and payroll payments for agricultural workers in Valle de Trinidad, in the north of the country.

In many cases, these were people who did not speak Spanish. We even developed SMS communication interfaces in Tarahumara to make the service accessible to them.

Bringing 7-Eleven into our operating network was another important part of the platform’s success. It meant that users could carry out transactions in the stores almost as if they were visiting a bank branch.

By 2014, the company had reached a valuation of $250 million in Silicon Valley, which was subsequently validated by Banco Santander Mexico and eventually led to an acquisition offer. That chapter came to an end in late 2017.

And then it was time to stop and think.

I needed time to step back, to understand how far we had come, and to figure out where we should go next.

For the first time in a long time, I had the luxury of not being in a hurry. I had time to reflect on what I wanted to do. And although I did not have a clear answer, I knew very clearly what I did not want to do.

For example, I knew I did not want to build another product aimed at end users.

I did not want to create a product with a specific set of functional capabilities for a particular market. I wanted to give myself the luxury of building something I was truly passionate about.

That may sound strange because, in many ways, I had already been doing what I loved. But, as I have always said: I am on an endless search for that indispensable something…

I needed to create technology.

Not management software. Not another platform designed to revolutionize a particular market.

No.

I needed to find the need. Find the problem. And have the vision to create the technology that could solve it.

I wanted to build technology that did not depend on the end consumer. I wanted to create a tool for engineers, something that could be used to build hundreds, even thousands, of different solutions.

Pure technology. Innovative technology. Technology that could earn us a tiny place in history, the space occupied by a single bit in the history of computing.

At the beginning of 2019, after participating as a partner in the launch of Spain’s first EDE, or Electronic Money Institution, I began thinking deeply about how software development had evolved toward microservices and API-based architectures, and how profoundly that was changing the industry.

But the most important part of those reflections was a much more fundamental question.

If APIs and microservices were becoming such an important part of the future of software development, how was it possible that there was no technology, and no programming language, specifically conceived, designed, and built for that purpose?

How could it be that we had HTML or Node for building websites; C and C++ for lower-level, machine-oriented development; and general-purpose languages such as Python and Java, but no programming language specifically designed for developing APIs?

But a question like that needed more than intuition. It needed to become technology. And to make that happen, I worked with Rafa Ruiz, who from the very beginning was instrumental in shaping and building what would eventually become AVAP. I had a clear understanding of the problem I wanted to solve and the need for a new programming paradigm; Rafa brought the knowledge and technical depth needed to turn that vision into a real architecture and language. Together, we began to shape AVAP.

That question was the seed of AVAP.

Then, in 2020, the COVID pandemic began, and we founded 101OBEX, now branded as AVAP.

That marked the beginning of a complex and completely unexpected chapter. We had designed the AVAP’s launch assuming relatively normal circumstances, and suddenly we found ourselves building the entire project in a completely new and extraordinary environment.

From 2020 until around 2023, when we returned to Palo Alto, our focus was essentially on designing the software and developing the first versions of the technology. During those years, Rafa and I worked closely on the architecture, language, and runtime that would ultimately define AVAP. It was a deeply technical process that required us to challenge many of the conventions on which software development had been built and to start designing from the ground up a technology conceived for a fundamentally different model.

At its core was the creation of the first virtual programming language designed for building microservices, which, at that time, had one primary technical application: APIs.

One of the most important moments in this entire process was the opportunity we had to immerse ourselves in Silicon Valley, to meet people, learn, explore, and work directly with OpenAI and different entrepreneurial communities.

That experience allowed us to bring together and consolidate our vision of AI-assisted and AI-driven software development, and to understand the impact it was going to have in the short and medium term.

Eventually, we became very clear about what we wanted to build.

We wanted to create technology that could enable the development of components containing business logic and manage their entire lifecycle, with the ultimate goal of reaching a future where AI could autonomously develop and operate those components, based on a defined set of functional specifications.

By 2023, Rafa and I had a very clear understanding of both our starting point and our long-term vision. We had built the technological foundation and were beginning to understand much more clearly just how far that architecture could go.

What remained completely uncertain was the path between the two.

And that was fine.

We knew we would have to navigate that path as new challenges emerged.

First, we needed the technology and the language.

Then, we needed the commercial products.

And finally, we needed the AI strategy, which, at that point, we simply were not ready to develop.

We developed the products and brought them together into the AVAP platform. We then created the SaaS version.

And in mid-2025, we launched Brunix, the AI layer that allows engineers and autonomous agents to operate the same live backend simultaneously, with complete cloud and edge deployment.

Brunix is not simply about providing knowledge for building and operating intelligent backends     . It also allows users of AI Systems to have their own AI, with knowledge of their specific system, both in the cloud and at the edge.

This creates something the market hasn’t solved: a bank, a hospital, or a government agency no longer has to choose between running modern AI infrastructure and keeping their data inside their own walls. On-prem is not a compromise. It is a first-class deployment.

And, in many ways, this brings the story back to where it started: with my passion for technology, my obsession with solving problems that do not yet have answers, and my desire to create technology that can become part of the history of computing.

2.     Please share the message, vision, and achievements you believe best define your leadership journey.

I believe the central theme starts with passion.

Passion is what pushes us to challenge ourselves, to look for more, and to constantly strive to become better. But passion is sustained by sacrifice, hard work, and knowledge.

Everything in life involves sacrifice. Throughout the day, we make hundreds of decisions, and every decision comes with one or more things we have to give up. Most of the time, those choices involve some form of sacrifice.

Hard and consistent work requires sacrifice. Perseverance requires sacrifice. And none of that is enough without knowledge.

But knowledge itself is not easy. Identifying what you need to know, pursuing it, acquiring it, maintaining it, keeping it up to date, and, most importantly, being able to turn that knowledge into something real and useful is not only difficult; it requires sacrifice.

The world is full of brilliant minds. It is full of people with great ideas and hundreds of valuable innovations that could be brought to life.

But ideas alone are not enough.

You need people who are willing to sacrifice many things to make any of those projects a reality. And you need the capabilities, the talent, and the determination to build something from the ground up. In my case, I have shared that journey especially with Rafa, with whom I have worked from the very conception of AVAP’s technology through the entire process of turning that vision into a real technology.

I think one of the achievements I value and enjoy the most is being able to work with a very small team of people whom I deeply admire, many of whom I have been working with since my days at Pademobile.

The people on my first line of leadership have been working with me for an average of 15 years.

That, for me, says a great deal.

It speaks to trust, loyalty, shared experience, and a common belief in what we are trying to build. We have gone through different companies, different countries, different challenges, successes, failures, and moments of uncertainty together.

And perhaps that is what I value most about leadership: not simply being the person at the top, but building a team willing to turn an idea into reality for the engineers who are building modern AI systems in 30-year-old languages. Not because they chose those tools, but because nobody had built better ones yet.

3.     How did your experience in building a platform that accelerates API development and lifecycle management pave the way for AVAP to become a go-to platform for building and operating AI systems?

Knowledge and experience are almost everything.

I have had the privilege of witnessing the birth of home computing from its very beginnings. I have worked in corporate environments with mainframes and personal computers. I have seen technologies emerge, evolve, become mainstream, and eventually disappear. I have experienced their entire lifecycle, and that gives me a very important perspective, not only on technology itself, but also on the business of technology and on how people interact with it.

On top of that, I have had the opportunity to work across different markets and with very different business cultures. That has allowed me to see different approaches, ways of thinking, and ways of solving essentially the same problems.

All of that experience has led me to look for and build solutions that are not designed around a single reality, but that can have a genuinely global perspective and application.

There is another dimension to that experience that I consider particularly important: I have also experienced technology from the developer’s perspective.

Very often, technological evolution and new tools are designed around what the system can do, what the business needs, or where the market is heading. But they are not necessarily designed around the person who is actually building with those tools.

As a developer, I have experienced that gap firsthand. And, to a large extent, it has shaped the way I think about innovation.

Technology should not only be technically advanced. It should be designed for the people who use it and should enable them to do their jobs better.

A significant part of the motivation behind AVAP comes precisely from that idea: creating technology that puts the developer back at the center, removes unnecessary complexity, and makes it possible to build systems in a more natural, dynamic, and efficient way.

That perspective is also what led to our evolution from APIs to AI Systems.

In 2019, we began questioning the traditional model. Instead of building yet another layer on top of conventional API infrastructure, we decided to create a programming language specifically designed for APIs: AVAP.

Over time, that same architecture allowed us to understand that the problem was much bigger than simply building and managing APIs.

As artificial intelligence and autonomous agents began interacting continuously with APIs, enterprise knowledge, and running systems, we realized that the technology we had originally created to make backends more dynamic, and to allow them to evolve without constantly having to rebuild and redeploy the entire system, was also the right foundation for building and operating AI Systems.

The fundamental problem was essentially the same.

What had changed was the world around it, and therefore the needs we had to solve.

I believe one of my greatest strengths is the passion I have for technology. Beyond having management, commercial, or marketing skills, I am genuinely passionate about technology.

And I believe having that foundation and truly understanding the technology you are building is incredibly valuable.

When you innovate, it is not only the vision of the future that matters. The ability to understand the past is equally important.

You need to be able to assess how long an innovation is likely to remain relevant, what its technical requirements will be, what it will cost, and what other factors will determine whether it can actually succeed.

There is much more involved than simply finding a solution to a problem.

You can have the perfect technical solution to a problem and still fail because it is technically unfeasible, economically unsustainable, or incompatible with the regulatory environment at that particular moment.

For me, that is an essential part of innovation: not just having the vision to see what could exist, but having enough experience to understand what can actually be built, adopted, sustained, and evolved over time.

4.     APIs have evolved from being integration tools to becoming the core infrastructure powering AI-native applications, autonomous agents, and intelligent digital experiences. Could you share your perspective on this evolution and explain where AVAP fits into it?

APIs were never simple. But they have always had the ability to look simple, and that is actually a very valuable quality.

The reality is that, even if we change the terminology, and, of course, there are attributes that make them different, MCP Servers and tools are, at their core, APIs.

So I agree with the statement: without APIs, there is no AI. Without APIs, there is nothing.

I also believe AI has accelerated the process by which many technologies become legacy. What was, until very recently, a fantastic business (API gateways, for example), I believe is already becoming a commodity, just like many other products that once played a very important role around the API lifecycle. The reason is simple: they were built as layers on top of existing infrastructure, rather than introducing a new language and programming model underneath.

What was, until very recently, a fantastic business- API gateways, for example- I believe is already becoming a commodity, just like many other products that once played a very important role around the API lifecycle.

I would still keep orchestrators outside that commodity category, but I don’t think it will be long before they get there as well.

Everything is changing constantly today, and it is changing at a dizzying speed.

And, by the way, I say this as someone whose company AVAP won the Best Innovation in API Gateway 2026 award from API World this year. The recognition follows Kong as the previous category leader, marking AVAP’s emergence as a new technology in the API ecosystem.

Technology has always evolved this way. So have programming languages and development paradigms.

Python and Java are powerful, proven languages. But they were designed before the problem we’re solving today existed. They weren’t built for an interpreted, agent-operable backend where humans and AI work on the same running system.

I would never be reckless enough to simply call Python or Java legacy, just as I would never call C legacy.

What I do believe is that C has a very clearly defined place in the world of software development. At the same time, there are languages that are used for almost everything, largely because they occupy a space that no specific-purpose language has yet filled.

And that is an important distinction.

This same evolution led us to understand that the real value was not simply in having created a virtual language for API development. The deeper value was in the technology underneath it.

That is where PLATON was born: an open-source virtual programming language kernel that enables the creation of purpose-specific virtual languages and makes them compatible with any product that supports the technology.

Because PLATON is open source, customers aren’t locked into a proprietary runtime they cannot inspect or audit. That distinction matters considerably to regulated enterprises.

AVAP was our first implementation and our proof that the model worked.

But our vision goes far beyond APIs.

What we are building is the ability to take this concept of virtual languages into any domain where there is a need to create, evolve, and operate software in a fundamentally different way.

APIs were the starting point. They were never the destination.

5.     Through the AVAP, organizations can create intelligent backends where humans and AI collaborate to build, operate, and continuously evolve enterprise software. Could you explain in detail how AVAP is redefining modern software development for the AI era?

In reality, the responsibility for this capability lies with Brunix, not with AVAP      itself as a platform or product suite.

Brunix is our AI, and it has several characteristics that make it very different.

To begin with, Brunix is not a simple LLM. It is an architecture, a RAG made up of hundreds of components, including several LLMs working together, dozens of agents, and other specialized components.

Brunix has been trained to deeply understand our technology. It knows PLATON, AVAP, every one of our products, and how they are operated. That makes it an extremely powerful programming assistant, code generator, implementer, and infrastructure administrator.

Because of all these capabilities, you can give Brunix a functional specification, and it can create the complete intelligent backend required, deploy it to production, and make it available for consumption.

And it can do this because 100% of the AVAP platform is exposed to our AI through internally managed MCP tools.

What makes Brunix particularly different, however, is the way it has been trained.

Most LLMs on the market that provide coding copilots for developers work from training based on millions of examples of source code. They retrieve or generate what they consider to be the most appropriate pieces and assemble what can sometimes look like a wonderful collage of code.

But they do not necessarily test that code before giving it to you.

Our approach with Brunix has been fundamentally different.

We have trained it to work as closely as possible to the way a programmer actually thinks: understanding how to program, how a system should be built, and, most importantly, why something needs to be built in the first place.

In fact, Brunix does not generate code directly. It generates AVAP ASTs (Abstract Syntax Trees), and then a deterministic layer generates the AVAP code from those ASTs.

That makes the resulting code much more closely aligned with the intended purpose of the request.

And there is another important difference: the code generated by Brunix is tested by the compiler before it is delivered to the developer.

That is something the AI coding tools currently available on the market generally do not guarantee.

With our AI, the code is tested before it reaches the programmer.

It works. Every time.

There is another important aspect to consider: Brunix’s architecture.

Brunix operates in the cloud, but it is built around a distributed knowledge architecture. This allows customers to have their own Brunix, with its own identity, and to manage the data and knowledge of their AI System either in the cloud or locally.

This is enabled through our entire edge strategy, powered by Brunix AIOS and our edge offering, built primarily around NVIDIA hardware, the manufacturer we work with as an official certified technology partner.

The ability to manage not only identity, context, and knowledge, but also to determine, simply and visually, the behavior and flows of an AI, is extremely powerful.

Through graphs within distributed intelligence and knowledge architectures, we can make changes and adaptations to the behavior of the AI in real time, without having to rebuild the underlying system.

And I believe this is one of the areas where Brunix brings significant value to the market.

We are not simply talking about an AI that helps you write code.

We are talking about an AI that understands the system, understands its knowledge and context, can build it, deploy it, operate it, and continuously adapt how it behaves.

That is a very different model of software development, and it is the direction in which we believe enterprise software is heading.

6.     In an era where AI companies are rapidly emerging, each claiming to outshine the rest, what sets your AVAP apart? What unique advantages do your solutions offer, and why should clients choose you over the competition?

 

It is true that today everything is AI, and that thousands of new AI solutions are emerging around the world every month.

But that is not new.

It has happened with every cycle of technological innovation.

Think about the number of e-commerce businesses that appeared during the dot-com bubble in the late 1990s. Suddenly, anyone could open an online store. But selling online also meant understanding logistics, at least to some degree, and understanding finance, at the very least, knowing the difference between a debit and a credit.

The same thing happened with other technological innovations in previous decades, and now it is happening with AI. New technology creates access to enormous opportunities, but that opportunity can sometimes make us overlook the fact that, beyond the innovation itself, there are still many things you need to know, understand, implement, and manage.

As a result, many projects that start as genuinely good ideas ultimately go nowhere. Others lose their value very quickly because they can be easily replicated, or because the core of the business depends on a third party that eventually dedicates the necessary time and applies the necessary knowledge to build the same capability.

That said, I think this behavior is part of the entrepreneurial model. It is part of what entrepreneurship is, and ultimately part of how people behave, especially when you are young and full of the desire to innovate and create.

And I think one of the biggest things happening right now is that we are confusing having an AI or having an AI solution with simply having a solution that consumes the API of ChatGPT, Claude, or another model.

That is extremely risky.

Because, ultimately, something like that can be replicated by almost anyone in a very short period of time, leaving you with a business that is completely dependent on someone else.

Most AI solutions today are wrappers around ChatGPT or Claude. That works for a demo. It is not a defensible business.

So yes, there are thousands of AI solutions.

Putting AI into a PowerPoint is easy.

Every major platform and application on the market is now “AI.”

And, for an entrepreneur with enough money, it is arguably even easier to change your branding, launch thousands of ads, go on the road to industry events, and announce that you are now an AI company.

There is some truth in all of this. But in many cases, there is a huge percentage of future truth, something that may eventually become true, and a small residual percentage of fiction.

The challenge is knowing how to tell the difference.

What we need to evaluate are the ideas that are technologically aligned with the innovation and that have genuine intellectual property value behind them.

That is why we are so committed to owning 100% of our technology.

I firmly believe that there is currently no single technology, platform, or product suite that can match what Brunix offers from a single technology provider.

You can assemble something that may look similar by combining different technologies and vendors, but the comparison becomes completely different when you consider the administrative cost, technical stack complexity, integration requirements, and, of course, the level of service.

But ultimately, the difference with AVAP is not that we took an existing technology and added AI to it.

The difference is that we built the technology around the problem we are solving today.

AVAP began as an interpreted virtual programming language designed to allow the backend to evolve without constantly going through compilation, rebuilding, and deployment cycles.

On that foundation, we built the AVAP technological platform. Development, execution, governance, knowledge, and operations coexist within the same architecture, allowing humans and AI agents to work on the same backend.

And this is particularly important because we are not talking about a collection of existing products that we have integrated in order to enter the AI market.

We are talking about an architecture conceived from the language level to build and operate AI Systems.

That, in my opinion, is our greatest advantage.

While others are adding AI capabilities to infrastructures that were designed for a different paradigm, we are building the infrastructure for the paradigm that comes next.

Explore Raúl Nogales’ “CEO of the Year” Story

AVAP – Pioneering a New Development Software Paradigm: AI Systems