Why Every Company Needs a New Enterprise AI Expert

The corporate race to dominate with artificial intelligence is shifting from broad adoption to deep specialization. For IT professionals wondering where to steer their careers, the most strategic move is to pursue the skills required for what may become the most critical role in enterprise AI: the frontier engineer.

Steve Lucas, CEO of integration firm Boomi, laid out this vision during a one-on-one conversation with ZDNET at the company’s World Tour event in London. He described the frontier engineer as a professional with advanced expertise in data science and neural networking who will be the linchpin for organizations aiming to unlock real competitive advantage through AI. “Companies will thrive when they truly grasp how to optimize frontier models and apply them daily,” Lucas said. “A CIO must have that person embedded in the organization.”

With a career spanning leadership roles at Salesforce, SAP, iCIMS, and Marketo, Lucas has a clear view of tech’s trajectory. He now helps Boomi and its clients navigate the complex era of AI agents, where data use is more critical than ever. Within this environment, the frontier engineer is poised to provide much-needed clarity. But this is no entry-level gig. The role demands rare, specialized qualifications and carries weighty responsibilities. As a result, these experts will be in fierce demand as companies scramble to outpace competitors.

Lucas posed a pointed question to every CEO: “Is there a single person in your company who understands how neural networks work? For 95% of organizations, the answer is no. And that is the fire we are playing with.”

To understand why this role matters, Lucas traced the evolution of hyped AI jobs. Prompt engineering was once the hot ticket, with endless advice on extracting answers from models. Then came the release of OpenClaw in late 2025, shifting the focus to harness engineering for operational software layers. More recently, loop engineering , designing feedback loops for AI coding agents , became a must-have skill. “Go to Indeed and type ‘loop engineer,’ and jobs pop up,” Lucas noted, shaking his head at the rapid churn. He compared these fleeting roles to “quarks and bosons that pop into existence and then disappear.”

The danger for IT professionals chasing these trends? A career dead end. “Those are not enduring skills,” Lucas warned. “Enduring skills are understanding data science and neural networking deeply.”

The frontier engineer role, by contrast, is built on lasting foundations. Lucas traced its origins to the demand for frontier model engineers at Big Tech firms pioneering AI. These individuals are exceptionally rare. “I would bet there are fewer than 3,000 people worldwide , maybe even fewer than 2,000 , who know how to build and train a model at scale, understand neural networks, and build a backpropagation-oriented large language model,” he said.

But a tipping point is near. More end-user organizations, not just tech giants, will need someone who can apply deep knowledge of frontier models to real-world problems. “Enterprises will need someone who really understands how neural nets work , not necessarily how to build them, but how to optimize them,” Lucas explained. At a minimum, this requires an advanced degree in data and neural networking. Combining these two domains is uncommon, especially in non-tech businesses. “It’s one thing to understand data,” Lucas said. “It’s another to understand neural networks. But applying that capability in the enterprise is profoundly important.”

So how does a frontier engineer differ from a chief AI officer? Lucas sees the CAIO as focused on models, frameworks, compliance, and organizational integration. The frontier engineer, in contrast, sits between the AI executive and the hands-on builder. Think of them as a specialized version of a forward-deployed engineer , a role championed by companies like Palantir , who iterates and applies solutions to business problems. The frontier engineer knows enough about the technology’s inner workings to squeeze every drop of productivity from models.

Finding these people is tough, even at Boomi. “It’s difficult to pin down the one person who is a deep expert on neural nets and can help me maximize output,” Lucas admitted. Yet for up-and-coming tech professionals, becoming that rare individual is a pathway to long-term success. “In the end, that capability will mean the difference between winning and losing.”