Training Models, Not Just Machines: Building Human-Centered AI That Listens, Sees, and Includes

In the race to build bigger models, we often forget the people on the other end of our systems—those who need AI not just to be powerful, but to be present. In this session, we'll explore how to build human-centered AI that listens to the unheard, sees the unseen, and includes the overlooked. You’ll get a practical, technical walkthrough of how to use Azure AI services (OpenAI, Speech, Vision, and Translator) to build intelligent applications that serve deaf, blind, and multilingual users. We’ll also look under the hood at the design patterns, accessibility challenges, and ethical considerations when deploying multimodal AI at scale. This is not just about prompt engineering — it’s about purposeful engineering. Whether you're a developer, ML engineer, or architect, you'll walk away with the tools and mindset to build AI that doesn’t just work, but works for everyone.