What is one hands-on lab you’ve run that teaches on-device edge AI using low-power NPUs on student laptops, inspired by recent CES trends? What result did you see in student performance or engagement?

I haven't run a specific hands-on lab for on-device edge AI with NPUs on student laptops, but I can tell you what works from my workshop experience. When I taught the 3-part AI workshop for the US Office of Foreign Broadcasting with journalists, the game-changer was letting them immediately apply AI tools to their actual work problems--not theoretical exercises. The key insight from that workshop: participants retained way more when they used AI to solve their real challenges in real-time. We had journalists using LLMs to research story angles and fact-check during the session, and the "aha moments" came when they saw how it augmented their skills rather than replaced them. Engagement shot up because it was their content, their problems. For students with laptops at CES-style events, I'd focus on computer vision tasks using their built-in cameras--like training a model to identify objects around the convention center or analyze crowd patterns. HP's Elite Dragonfly Max that I covered at CES 2021 had AI-driven features built in, and that's the kind of hardware students already have access to. The hands-on approach always wins--it's why I love workshops over lectures.

One hands-on lab I've run that teaches on-device edge AI using low-power NPUs was inspired directly by what we're seeing at CES around private, real-time AI on consumer hardware. In this lab, students used their own laptops to deploy a lightweight gut-health symptom classifier that ran entirely on the device—no cloud calls—using the built-in NPU for inference. We simulated real clinical constraints by feeding it anonymized symptom patterns and requiring millisecond-level responses while keeping battery drain minimal, mirroring how future health apps will function in the real world. What stood out immediately was how engagement shifted once students saw AI working offline, instantly, and securely. One student told me it was the first time AI felt "real" rather than theoretical, because the model responded faster than a cloud API and protected patient privacy by design. Performance improved as well—students who had struggled with abstract ML concepts were suddenly optimizing model size, power usage, and inference speed with confidence. That hands-on exposure made edge AI tangible, and it reinforced a critical lesson: the future of AI in healthcare and beyond isn't just smarter models, but smarter deployment where efficiency, privacy, and real-world constraints matter.

The hands-on lab focuses on teaching students about on-device edge AI using low-power Neural Processing Units (NPUs). Students use their laptops to create and deploy real-time AI applications, gaining practical experience without cloud reliance. The curriculum covers AI fundamentals and emphasizes energy-efficient edge computing, drawing inspiration from trends at CES. Overall, the lab aims to provide foundational knowledge and practical skills in edge AI technologies.

I'm a landscaping company owner, not a tech instructor, so this question isn't in my wheelhouse. But I can share what I've learned about hands-on training from running equipment safety workshops for our crew at Lawn Care Plus. When we train team members on new commercial mowers or hardscaping tools, the retention rate jumps when they operate the equipment on an actual job site within the first day--not just watch demos. We had a new hire struggle with our paver installation techniques until we let him lead a small walkway project in Roslindale. His confidence and speed improved by roughly 40% compared to guys who only shadowed for their first week. The parallel to your AI question: people learn by doing with real stakes, not simulations. If I were setting up a student lab, I'd have them solve a tangible problem their campus faces--like optimizing energy usage in a dorm using on-device processing. Give them something where failure means they see actual consequences, and success means they helped someone. Make the lab about a result they can point to and say "I built that," not just a grade. That's what keeps people engaged, whether it's landscaping or AI.