How Alister Mac Cormack Turned Self-Taught Engineering Into a Career Deploying AI

At 12 years old, Alister Mac Cormack wanted a computer that could run the video games his existing machine could barely handle. Buying one was out of reach, so he decided to build it himself. The technical challenge was only part of the problem. He also needed to earn enough money to pay for the components.

Mac Cormack began saving and taking on work for other people. Those early jobs required him to listen carefully, understand what someone needed, and produce something useful. The experience connected technical curiosity with the practical demands of solving a problem for another person.

"Building the computer taught me about the hardware," Mac Cormack said. "Earning the money for it taught me how to understand what someone needed and deliver work that helped them."

Once the machine was assembled, new questions replaced the old ones. Mac Cormack knew what the parts did, but he wanted to understand how software controlled them. A programming book from his school library gave him a place to begin.

Learning to Build Without a Set Path

Mac Cormack taught himself to code as a teenager by working through the book and testing what he learned. Programming let him move beyond using computers and begin shaping what they could do.

"The first time I sat behind a computer, it felt like magic," he said. "Coding was how I started understanding what was happening underneath and how I could build something myself."

That interest soon moved beyond personal projects. Mac Cormack began solving problems for other people and working with startups in different industries. His experience included medical systems and residential solar, followed by work in banking and cybersecurity.

Each setting gave him a new kind of problem to examine. The subject matter changed, but the process remained similar. He had to understand an unfamiliar system, identify what was failing or missing, and determine whether software could improve it.

Mac Cormack continued working while completing a demanding, in-person university program. The schedule required him to negotiate unconventional arrangements with teachers and employers. Many days ended with more work after classes were finished.

"In Chile, most people do not start working in the field until their final year or after graduation," he said. "I wanted to start earlier, so I had to look for opportunities and convince people that I could do the work."

Letting Results Establish Credibility

Mac Cormack often entered professional settings without the credentials employers expected to see. Some underestimated him before reviewing what he had already built or solved.

He responded by focusing on the work itself.

"I had to show through results what a diploma is often assumed to show," he said. "That pushed me to be direct about what I could do and to prove it through the outcome."

His interest in difficult systems eventually led him deeper into cybersecurity. Mac Cormack performed penetration testing for banks and other financial entities while working at Nivel4. He also conducted independent penetration testing for the Chilean Ministry of Justice.

Security work required him to study software from the perspective of someone trying to make it fail. That experience changed the way he thought about engineering.

"Breaking a system is like building it in reverse," Mac Cormack said. "You look at the assumptions behind it and find out where they stop holding up. That makes you more careful when you build."

His research led to CVE identifiers being assigned under his name for vulnerabilities in open-source software. In 2024, the Chilean government's Computer Security Incident Response Team included him in its Hall of Fame for responsible disclosure.

Mac Cormack also co-founded Security UC Club and organized Hack.ING, a university hacking competition with approximately 270 participants from Chile, Peru, and Colombia. The event gave students a setting where they could test their skills against unfamiliar security problems.

A Role That Requires More Than Installation

Mac Cormack's work across startups, software engineering, and cybersecurity eventually led him to forward-deployed engineering. The role places an engineer close to the company using the technology rather than separating product development from day-to-day business operations.

He now works from Chile as a forward-deployed engineer at Cognition, where he was the first person hired into the position from Chile. His job begins with learning how a company functions before deciding what an AI system should do inside it.

"The first step is understanding the company," Mac Cormack said. "You have to learn the systems people use, the data involved, and the actual problem. The technology comes after that."

Some assignments may involve installing or adapting an existing product. Others require custom engineering because standard software does not fit the company's processes or technical environment.

Mac Cormack sees that gap between a product demonstration and a dependable deployment as one of the most important parts of the role. Software that performs well under controlled conditions may still need substantial work before employees can rely on it within existing systems.

"An off-the-shelf product may solve part of the problem," he said. "The harder work is making it fit the environment where people will use it every day."

His cybersecurity background also affects how he approaches AI deployments connected to company data. Technical usefulness is only one consideration. The system must also be evaluated for the risks created by how it connects with internal software and information.

Why Range Matters in Applied AI

Mac Cormack believes engineers benefit from exposure to industries that appear unrelated. Knowledge developed in one setting can reveal a solution or risk in another.

His own path has moved through several kinds of technical work rather than one narrow specialty. That range helps him recognize patterns across business systems while still examining the engineering details closely.

Mac Cormack earned first place at Platanus Hacks in 2024 and second place in 2025. He was also selected for Y Combinator's first AI Startup School in 2025. [Project names, team structure, and Mac Cormack's specific contribution should be added if confirmed.]

He does not view self-teaching as an easier alternative to formal education. His experience required long hours of practice and repeated attempts that produced no immediate recognition.

"Working smart matters, but it does not replace the hours," Mac Cormack said. "The repetitions are what build judgment. You learn how to handle the next difficult problem because you have already spent time working through the last one."

Forward-deployed engineering keeps him close to that kind of challenge. He enters companies where the systems are unfamiliar, and the answer is not already packaged.

For Alister Mac Cormack, the work begins by taking the problem apart carefully enough to understand what should be built next.