Proving the art of the possible: AWS prototyping engineer Patrick O’Connor

Patrick O’Connor, EWMBA 27, is a senior AI prototyping engineer at Amazon Web Services, where he leads applied AI prototyping with customers and drives some of Amazon’s most ambitious agentic AI initiatives. In this first person perspective, he traces his path from Ireland and Australia to AWS—and how the connections he’s making at UC Berkeley Haas are pushing his career into new territory.

“I grew up between worlds. With an Irish dad and a Chilean mum, I spent my first 11 years in Ireland followed by 12 in Australia.

Wherever we lived, technology was the constant, and it found me through my father. He’d been in tech since college, starting in networking before moving into radio-frequency identification (RFID). It was during his years leading FE Technologies, a library RFID company, that I got my first taste of programming. From the physical RFID kiosks to the software behind them, I started building young. That grew into a natural interest in computer science, and Monash University became the scaling factor. At one of the largest engineering schools in Australia, I undertook a double degree in commerce and computer science, accounting on one side and advanced computer science on the other.

In my final year, 2020, as COVID swept the globe and Australia entered some of the world’s strictest restrictions, I was lucky enough to land an industry-based learning placement at the National Australia Bank. Demand for engineers exploded, far outstripping supply, opening up a myriad of opportunities for me to learn and lead.

It was at the bank that I first encountered AWS. One day my manager asked me to ‘ssh into an EC2’, leaving me confused, not just by the lingo but by what any of it had to do with Amazon, which I’d always thought of as an e-commerce site.

That moment became the catalyst for a whole new realm of discovery. The deeper I dove, the more I saw just how critical the cloud is to a connected world: AWS alone has more than 240 services powering much of what’s online. There was no quenching my curiosity; all I wanted to do was build cool things. Eventually that curiosity carried me to Amazon itself. I joined AWS in March 2021 as one of the first AI engineering hires in the professional services AI/ML practice.

Joining AWS’s Prototyping Team 

Professional services showed me what proven technology could deliver. What pulled me to the specialist prototyping team was the opposite: an appetite for pure innovation, for pushing on what technology can’t quite do yet.

When people hear ‘prototyping’, they usually picture a quick demo, something to show short term value and get a conversation rolling. What we do is different. Prototyping means running experiments with customers to prove out the art of the possible: taking the latest developments in technology, combining them with emerging tech and research, and pointing all of it at a genuine business challenge. It sits between a proof of concept and a pilot: real enough to prove the idea works, early enough that failure is still acceptable. And it’s intense. In the span of four to six weeks, a prototyping engineer wears every hat there is, from builder and coder to consultant and evaluator, all in the hope of pushing technology far enough to unblock an entire industry, not just a single customer.

That mandate has taken me to some unlikely places.

With HERE Technologies, the global leader in mapping, we took on one of the most tedious jobs in autonomous driving. Before a self-driving system can be trusted, it has to be tested against real world road scenarios and engineers were manually hunting through map data to find them. We built a tool that lets them describe what they need in plain English, like a highway off ramp at a particular incline or a complex intersection and get back a simulation-ready scene. HERE launched it as a product called SceneXtract, and it draws on the same mapping data behind systems like Mercedes-Benz’s Drive Pilot. That’s the outcome you hope for: the prototype doesn’t stay a prototype.

And with Metagenomi, a gene editing company, the problem was pure scale. Somewhere among the billions of proteins encoded across the tree of life are enzymes that could become the next CRISPR, but finding them means searching a database of 3.5 billion candidates. We turned that hunt into something closer to a search engine: instead of scanning sequence by sequence, their scientists can now find the closest matches to a promising protein in seconds, for a fraction of a cent per query. It’s already helped them discover novel enzyme families they hadn’t been able to reach before.

Why an MBA, Why Now 

In the middle of all this building, I decided to get an MBA, for two reasons. The first is ambition. I have a long term vision of becoming a technical leader in an organization, or one day a founder. The MBA equips me for that: it complements my engineering skillset, challenges me on things I don’t typically encounter in my line of work and surrounds me with exceptional people I learn from constantly, all while growing alongside a job I love. The second reason is simpler. I have the time. This is the season of my life when I can pour myself into something this intensive and there will never be a better moment to invest in it.

I chose Berkeley because the depth of its ecosystem is unmatched, as is its proximity to San Francisco, the epicenter of AI. But what has shaped me most so far is the faculty. Working in tech, you move at a relentless pace; there’s always a next build, a next release. Haas has given me professors who stretch your horizon beyond that rhythm: world class experts who take the time to challenge you and set you up for success. In Abhishek Nagaraj’s AI for Business Leaders, we examine how leaders should operate in a rapidly changing AI world, drawing on the paradigms of earlier general purpose technologies, like how the chip revolutionized computing and applying those lessons to navigate the moment we’re in now. And over the summer I took Frontier Ventures: Space, Air, Marine & Arctic with Olaf Groth and Adair Morse, a class that pushes you to think about entirely new economic frontiers.

Learning to Build Bigger 

It was through Frontier Ventures that Professors Groth and Morse connected me with Justin Wiley, who oversees a $25 million research portfolio at UC Berkeley’s Institute of Transportation Studies. Justin organized a visit to the Richmond Field Station. We toured the facility, met early stage startups working on physical AI, energy, and transportation along with Berkeley researchers. Among them was Professor Alexandre Bayen, who is working on coordinating safe landings for a future filled with eVTOL aircraft: vehicles that take off and land vertically like helicopters, but promise greener, quieter urban flight. These are precisely the kinds of problems I gravitate toward: novel, hard and blocking an entire industry. I’m now exploring how the work I do in prototyping could connect with initiatives like the Field Station.

The longer I’m at Haas, the more I realize that business is its own kind of engineering: strategy, storytelling, the long arc of a decision. I had only a primitive grasp of it before. That’s what’s really changing me here. It’s pulling me out of pure engineering mode and saying, ‘Patrick, focus on the big picture’. Prototyping taught me how to build ambitious things fast. Haas is teaching me to build them further into the future.