As AI Blurs the Line Between Designer and Engineer, Hybrid Roles Emerge

The definition of good design is changing. For years, a product's value was measured by visual craft and an elegant user interface. But in the age of AI, the focus is less on appearance and more on a tool's functionality. As engineering fluency becomes a core competency for designers, organizations are beginning to rethink how they build products, hire talent, and measure success.

Experts like Eros Marcello, a Senior Generative and Agentic AI Consultant at Eli Lilly and Company, are putting this new theory into practice. A multidisciplinary product designer with a decade of experience specializing in human-AI interaction, Marcello has a solid track record of shipping consumer and enterprise-grade AI systems for high-profile companies, including Meta, Apple, Salesforce, and Samsung. Today, his work focuses on developing hybrid workflows, orchestrating AI collaboration, and developing multi-modal interfaces.

"When it comes to AI, design is how it works, not so much how it looks," Marcello says. "It's very Steve Jobs-ian." Now, enterprises must move beyond the obsession with aesthetics and embrace a more rigorous, technical approach to design, he explains. The traditional separation between visual design and engineering simply no longer applies in today's AI-powered world.

According to Marcello, his perspective started to take shape nearly a decade ago. While working on Samsung's AI assistant, Bixby, he noticed a disconnect between the teams writing dialogue and the engineers building the underlying intelligence. "I immediately saw that what these systems say is marginal compared to what they do," he says. "We want them to act on our behalf with little to no oversight, not just respond or understand."

• Faking fluency: For Marcello, it was a realization that would eventually become the foundation of his core thesis. "If you don't understand the foundation of AI, you can't set expectations. The whole idea of being a non-technical AI/ML expert is illusory."

But engineering fluency also serves a much grander vision for technology’s role in our lives, Marcello explains. Whether we like it or not, AI is already an intermediary to the seamless future that we all secretly desire.

• Omnipotent operations: The future is a "love letter to ambient and ubiquitous computing, or what I prefer to call 'omnipotent computing,'" Marcello continues. "Apps have stagnated. Offline is the new luxury. We want less screen time and more autonomous experiences, which require interfaces to be, if not ambient, then completely invisible."

• Designing destiny: Marcello envisions the "mitigation of fragmentation" through multi-sensory, distributed intelligence—all driven by a simple, human-centric goal. "You want things that are precognitive and so radically intuitive and knowledgeable about you that they can perform things before you even know you have to do them. We want the minimal amount of taxation on your cognitive load. That is where I think true design lives."

Yet any vision of an "omnipotent" future stands in opposition to the way organizations have historically implemented AI. Rather than the models themselves, Marcello attributes early failures to the flawed environments in which they attempted to operate.

• Bandage blunders: Instead of a bandage for poor UX, Marcello says the intended purpose for AI should be to grant users entirely new capabilities. For example, he recalls consulting for a university where their new chatbot's primary purpose was to explain the "horrific" student web portal. His conclusion: "You're cannibalizing everything with this addition."

Successfully bridging this "literacy gap" requires a foundational understanding of failure modes like "hallucinations, adversarial prompt injections, vector embedding issues, or a simple API timeout," Marcello explains. But that doesn't mean every designer must become a full-stack ML engineer.

• Designer as diplomat: "You can make a chatbot that isn’t user-friendly, and you can make a super user-friendly UI for a chatbot that can’t function," he continues. "Why not put an ambassador in the middle between engineering, development, and creative to get both sides of the spectrum squared away?"

For Marcello, understanding these failure modes is what enables the rise of a new, hybrid role: the AI Design Engineer. This new role, he believes, requires a fundamental rethinking of how companies approach hiring.

• Function over form: The dysfunction has roots in how talent is evaluated, Marcello explains. "No one cares how beautiful your UI is anymore. Look at Facebook. Look at X. It's about function over form."

• Dashboard confessional: Today, the same trend emerges in the enterprise world, where pragmatism is often discouraged. "It's treated as a sin to say it, but what if the existing dashboard is the best solution? It works, and it's not broken."

• Coding catastrophes: Then Marcello takes aim at a system that rewards superficiality: "Vibe coding" with low-code tools is "causing catastrophe across enterprises," he explains. Because the tools are "formulaic," they cannot produce novel work.

The ultimate irony is that the technology intended to solve these problems has, in many ways, only created new ones. "I hoped AI would help, but it just made it all worse," Marcello says. "Now you have AI-generated resumes from fake candidates going into the ingestion engines of applicant tracking systems. It's worse than it's ever been."

In closing, Marcello reminds us that there is no "silver bullet" to fix this dysfunction. "The whole thing is sort of an illusion. It requires a major, system-level overhaul," he says. "Never mind 'what should we look for,' because depending on who you ask, you'll find a hundred different things, and 99 of them are going to be wrong. You need ubiquity. Everyone in the org needs to be on the same page."