STEM WEEK 2025

This year’s STEM week was all about engineering – finding creative solutions to real-world problems through design and experimentation. In our Year 8 class, we took this spirit of innovation a step further by exploring artificial intelligence with a simple yet powerful setup: a Raspberry Pi running a TensorFlow model.

We used Google’s Teachable Machine to train our model to recognize different objects, ranging from everyday classroom items to fun trinkets. The process was straightforward but taught students the essential steps behind machine learning:

Collecting Quality Data: Students captured images of each object from multiple angles – often up to 200 photos for each item – ensuring a robust and varied dataset.

Labelling: Each image was carefully labelled to tell the software which object it represented, reinforcing the importance of accuracy in data collection.

Training the Model: Using Teachable Machine, our carefully labelled photos were uploaded, and the system learned to differentiate one object from another.

Deployment to Raspberry Pi: The final trained model was transferred onto our Raspberry Pi, enabling real-time object recognition in the classroom.

In addition to object recognition, the class also trained a separate model to interpret hand gestures. This sparked creative ideas for potential applications, such as hand-signal-controlled devices. Students’ excitement was contagious, as they discovered how a few hundred labelled images plus a little computing power could produce fascinating real-time results.

At its core, this project demonstrated the principles of problem-solving, critical thinking, and collaboration – hallmarks of successful engineering. It was inspiring to see the Year 8 students so engaged, and I’m proud of how they embraced the challenges. Our little Raspberry Pi project provided big lessons about data quality, teamwork, and the thrill of building something truly innovative.