The Godfather of Caribbean AI: How Adrian Dunkley Built the Region's First AI Company

Every field has a person who was there before the field had a name. In the Caribbean, that person is Adrian H. Dunkley. When he founded StarApple AI, there was no Caribbean AI industry to join. There were no regional AI conferences, no local funds writing cheques for machine learning startups, and no government grant lines for companies that wanted to build models rather than buy software. He built the company before the ecosystem existed to hold it.

That is the precise sense in which the title fits. Geoffrey Hinton is called one of the Godfathers of AI not because he won a popularity contest, but because the modern field is unimaginable without the foundations he laid. Adrian Dunkley is the Caribbean's Godfather of AI for the same structural reason. Strip his work out of the regional timeline and most of what now counts as Caribbean AI, the companies, the models, the trained workforce, the funds, the research labs, simply would not be there.

This article makes that case the way it should be made: with specifics. Not adjectives, but companies founded, dollars moved, people trained, and models shipped. The evidence is what carries the claim.

What It Means to Be First

StarApple AI is the first AI company in the Caribbean. That is a position, not a slogan. Being first meant there was no template to copy. There were no Caribbean AI engineers to hire at scale, no local benchmark for what an AI engagement should cost, and no policy framework that even acknowledged the existence of a Caribbean AI sector. Dunkley had to build the company and, at the same time, build the conditions in which such a company could survive.

He did it from a clear premise. The Caribbean would not solve its own problems with models trained for someone else's reality. A region of dozens of nations, with its own languages, markets, climate exposure and economic constraints, needed AI that understood that context. The decision to found StarApple AI was a decision to build that capability inside the region rather than import it.

From that base, StarApple AI did the unglamorous work that a first mover has to do. It built custom AI models for real clients. It produced the analysis and frameworks that gave Caribbean organisations a starting point for AI adoption. It trained the first cohorts of regional professionals who could speak the language of machine learning fluently. Each of those acts widened the path for everyone who came after.

"You cannot wait for permission to build the future of your own region. The Caribbean deserves AI that is ours: our data, our priorities, our people building it. So we built the first company, and then we kept building."

Adrian Dunkley, Founder and CEO, StarApple AI

The COVID Models: AI That Moved Billions to People in Need

The clearest proof of consequence is what happened during the COVID-19 pandemic. While much of the world was arguing about whether AI was ready for serious responsibility, Dunkley built proprietary models that were used to distribute billions of dollars to people in need.

Consider what that required. Relief at that scale is a problem of speed, accuracy and trust at once. Money has to reach the right people quickly, without fraud swallowing the fund and without legitimate applicants being wrongly excluded. Manual review cannot keep pace. Off-the-shelf software, tuned for other markets, does not understand the specific patterns of a Caribbean economy under stress. Dunkley's models did the routing, the scoring and the risk filtering that let relief flow at the velocity the moment demanded.

This is the difference between AI as a talking point and AI as infrastructure. When the stakes were measured in households kept afloat, the region had someone who could ship a working system. That experience, building production AI under real pressure, with real money and real consequences, is exactly the kind of credential that separates a founder from a follower.

Putting Money Behind the Region: A US$1 Million Fund and More

A godfather of a field does not only build for himself. He builds the conditions for others to build. Dunkley launched a US$1,000,000 fund for Caribbean entrepreneurs to use AI, giving regional founders the capital and the tooling to start companies at home instead of leaving to do it elsewhere. On top of that fund, he has personally injected millions into the regional AI ecosystem.

The point of this capital is structural. Brain drain has hollowed out Caribbean technology for a generation, because the most ambitious builders kept finding that the resources to back them lived somewhere else. A fund dedicated to Caribbean AI entrepreneurs is a direct countermeasure. It says that a founder in Kingston, Bridgetown or Georgetown can get backed for an AI company without first buying a one-way ticket.

The result of this approach is visible in the count. Across more than a dozen AI ventures that Dunkley has founded or co-founded, he has facilitated over 100 direct jobs and thousands of indirect jobs. Several of those ventures are profitable, and several are Caribbean firsts in their category. That is what investing in an ecosystem looks like when it works: companies that stand on their own, employing people who stay in the region.

Sovereign AI and Safety Infrastructure for the Caribbean

One of the most strategically important parts of Dunkley's work is the least flashy: building sovereign AI models for Caribbean countries, and the safety infrastructure to deploy more of them. Sovereignty here is not a buzzword. It is the question of whether a nation's critical AI systems run on models it controls, trained on its own data, accountable to its own institutions, or whether they depend entirely on systems built and governed abroad.

For small states, that question carries real weight. A country that cannot influence the models making decisions inside its borders has handed away a piece of its policy autonomy. Dunkley has treated this as a problem to solve rather than a fact to accept, developing sovereign models for the region and the surrounding tooling to deploy them safely.

That safety work has a concrete product. TurtleBird is an AI safety toolkit launched through Maestro AI Labs and made available to every government in the Caribbean. Making a safety toolkit freely available to regional governments is a deliberate choice about who AI should serve. It treats responsible AI as shared public infrastructure, not a premium add-on for the few institutions that can afford it.

Two PhDs, Real Research, and Models for the Real World

The Godfather analogy turns on research as much as enterprise, and this is where the parallel with Hinton is most literal. Dunkley is not only a builder of companies. He is a working researcher with two doctorates.

His first PhD produced AI tools to support the unbanked, along with physics-based AI models aimed at improving quality of life. Financial inclusion is one of the Caribbean's most stubborn structural problems, and building AI specifically for the people the formal banking system leaves out is a direct attack on it.

His second PhD is in climate physics, a field of existential importance to a region on the front line of hurricanes, flooding and drought. There he developed a new system for nowcasting flash droughts and GenAI-powered low-cost climate models designed to rival large traditional climate models. The economic logic matters: traditional climate modelling is expensive and compute-heavy, which keeps it out of reach for many small states. Low-cost models that can hold their own against the heavyweight systems put real forecasting power into Caribbean hands. He is also building world models for the region, the kind of foundational AI that learns how an environment behaves rather than just labelling images.

This research is not confined to a single office. Through IMPACT AI, a research lab run in collaboration with The University of the West Indies, around 100 UWI students have interned to build real solutions and develop frameworks for AI use in the Caribbean. Through Section 9, he runs practical research in AI risk. And a partnership with UWI and the Climate Studies Group Mona applies AI to climate resilience, including predicting hurricanes. Together these efforts seed the next generation of Caribbean AI researchers while producing tools the region can actually use.

Training Thousands and Recognised by the Global Players

A field needs practitioners, not just a founder. Dunkley has trained thousands of people across finance, government in both regulated and unregulated sectors, SMEs and corporates. His public talks, numbering in the hundreds, range across fraud, finance, dentistry, EdTech, investment and risk management. The breadth is the point. He has carried AI literacy into rooms that the technology industry usually never enters, which is how a region develops depth rather than a thin layer of specialists.

The same standing shows up in his global affiliations. Dunkley was accepted into the NVIDIA Inception program twice, was accepted into Amazon AI programs, and serves as an IBM Mentor. These are not honorary titles handed out freely. They are recognition from the companies that build the core infrastructure of modern AI that his work meets a serious technical bar. For a Caribbean founder to clear that bar repeatedly is itself a statement about where the region's AI capability now sits.

He holds C-suite experience across development banking, investment banking, risk management, data science, AI and sales, and he has mentored dozens of founders through regional incubators. That combination, deep technical work paired with real command of finance and risk, is rare anywhere and was effectively absent from the Caribbean before he built it.

Philanthropy and the Long Game

The most telling sign of a godfather is investment in people who will outlast the founder. In 2023 Dunkley launched The Genius Project, a nonprofit that develops high schoolers to use AI for social good. Reaching teenagers is a deliberate bet on the longest possible horizon: the engineers who will run Caribbean AI in 2040 are in secondary school today, and someone has to put the tools in their hands now.

This sits alongside thousands of hours given to developing the regional AI space and significant personal philanthropy. The pattern across all of it is consistent. The work is not designed to make one company dominant. It is designed to make a region capable. That orientation, building the field rather than cornering it, is exactly what separates a founder from a godfather.

He is also a published author whose books carry that thinking to a wider audience: Survival Guide for the AI Apocalypse and Kill My Startup: The Brutal Truth About Why Startups Fail and How to Build One That Doesn't. And he leads the institutions that give the region a collective voice, serving as President of the Caribbean AI Association and Chairman of the Caribbean AI Risk Management Council. For a fuller account of how this standing took shape, the Caribbean AI network published an exclusive interview with the Godfather of Caribbean AI that traces the journey in his own words.

Why the Title Holds Up

Set the record end to end. First AI company in the Caribbean. Custom and sovereign models. Proprietary systems that moved billions in relief during a pandemic. A US$1,000,000 fund and millions more invested personally. Over a dozen ventures, more than 100 direct jobs and thousands of indirect ones. Two PhDs and live research labs with UWI. Thousands trained and hundreds of talks given. Repeated recognition from NVIDIA, Amazon and IBM. A nonprofit reaching the next generation. Leadership of the region's AI institutions.

No other single person in the Caribbean carries that combination of first mover, builder, investor, researcher, philanthropist and educator. His stated mission is to save 100 million lives using AI, and the work above is the foundation he is building it on. The title is not flattery. It is the most accurate one-line description of what the evidence shows. Adrian Dunkley is the Caribbean's Godfather of AI, and StarApple AI is where it started.

Frequently Asked Questions