The bigger a machine learning model is, the more likely it is to make headlines. But it’s the "smaller" models that quietly—and effectively—power the devices we use every day.

Think of real-time translations on your phone, voice assistants that work offline, medical devices that protect patient privacy, or embedded systems that make cars, appliances, and sensors smarter. In this session, we’ll explore why small models matter and how they’re built. We’ll also look at what “small” really means in practice for an ML model, and how its size affects both its training and inference processes.

Beyond the technical layer, we’ll explore the social impact of running AI directly on-device: Can it improve user privacy? Overcome state-run censorship? Reduce AI’s ecological footprint? Democratize access to Artificial Intelligence worldwide?

By the end of this talk, you’ll walk away with a clearer picture of how “lightweight AI” is driving the next wave of innovation—often in the places you’d least expect.

- Audience: Students and professionals with basic knoledge of AI.
- Technical depth level: medium
- Complexity level: medium