TeachingAItotheNextGenerationofEngineers

Teaching AI at university level while simultaneously leading technology teams in industry creates a productive tension. Every semester, I walk into a classroom at Universidad San Jorge knowing that some of what I taught last year is already outdated — and that my students will face an industry that looks different from the one I started in.

The Curriculum Dilemma

How do you teach a field that moves this fast? My approach has been to anchor the curriculum in fundamentals that don’t change — statistical reasoning, algorithmic thinking, experimental design, ethical frameworks — while keeping the tools and applications current. Students learn scikit-learn, and we explore how frameworks like TensorFlow or PyTorch fit into broader ML landscapes, but more importantly, they learn how to evaluate when to use which approach and why.

The hardest thing to teach isn’t the math or the code. It’s judgment. When should you use a simple model versus a complex one? When is more data actually helpful versus misleading? When should you deploy and when should you wait? How do you build systems that domain experts will actually understand and trust? These questions don’t have textbook answers, and they’re exactly the questions that matter most in industry.

What Students Teach Me

The best part of teaching is that students challenge assumptions I didn’t know I had. A student once asked me why we always frame AI success in terms of accuracy metrics, when the system she was building for a local nonprofit needed to optimize for fairness across demographic groups, even at the cost of overall accuracy. That question reshaped how I teach model evaluation.

Another student, working on a thesis about AI in sports coaching, pointed out that every paper we read assumed the coach was a passive recipient of AI recommendations — when in reality, coaches are experts who use AI as one input among many. That observation directly influenced how I design user interfaces for the athlete monitoring systems at RFEJYDA.

Bridging Academia and Industry

The most valuable thing I bring to the classroom is real-world context. When I explain the importance of careful API design and component architecture, I draw from my experience at IdoniaHealth as Frontend Tech Lead on CDTI medical imaging projects. Building user interfaces for healthcare AI teaches you quickly that a technically perfect backend is useless if clinicians can’t navigate it or understand what the system is showing them. When I discuss project management, I draw from my experience managing EU-funded R&D projects where the gap between a research prototype and a deployable system is measured in months and difficult conversations.

This bridge between academia and industry is what I find most fulfilling about teaching. It keeps me honest about what actually matters and ensures that the next generation of AI engineers enters the field with both technical depth, strong architectural thinking, and practical wisdom about how systems actually get used.