Multimodal GenAI in Action: Bridging Text, Vision, and Beyond

Join this insightful session to explore the exciting world of Multimodal Generative AI and its real-world impact. Discover how these powerful systems combine text, images, audio, and video to deliver richer, more human-like interactions. We’ll break down core architectures, explore alignment techniques, and showcase practical applications. Dive into two innovative systems: LLaVA, a vision-language assistant for visual Q&A, and AI Guide Dog (AIGD), which helps visually impaired users navigate in real time. Whether you're an AI enthusiast or a tech professional, this session will equip you with actionable insights into the future of multimodal AI.

Key Takeaways:

• Understand how multimodal Generative AI integrates text, images, audio, and video for richer interactions.
• Explore the core architectures and techniques that align multiple data modalities effectively.
• Discover real-world applications of multimodal GenAI in healthcare, entertainment, and navigation.
• Gain insights into systems like LLaVA and AI Guide Dog, showcasing practical multimodal AI implementations.

Aishwarya Jadhav is a Machine Learning Engineer at Waymo (Google), specializing in Perception systems for autonomous robo-taxis. Previously, she worked on Tesla's Autopilot team, contributing to AI models for Full Self-Driving and the Optimus Robot. With expertise in computer vision and large-scale ML, Aishwarya has also worked at Google Ads and Morgan Stanley. She holds a Master’s in Computer and Data Science from Carnegie Mellon University, where she led the AI Guide Dog project, developing real-time navigation systems for the visually impaired. You can reach her on LinkedIn.