The March BangPypers meetup, in collaboration with Google for Developers, provided valuable hands-on learning in the field of AI. Participants engaged in a workshop to develop a YouTube summarizer using Google Cloud and the google-genai library. The event also featured an insightful session on multi-modal models, offering practical knowledge and theoretical understanding.

Nikhil Rana expertly guided attendees through setting up their Google Cloud consoles and demonstrated the process of building GenAI-powered applications using Vertex AI. He then showcased a live YouTube summarizer application running on Google Cloud CodeRun.

Following the workshop, Abhik Sarkar provided a comprehensive overview of multi-modal models, emphasizing their ability to integrate heterogeneous data sources, specifically text and images. He began by outlining the evolution from single-modality AI, highlighting deep learning’s advancements in individual text and image processing. He then detailed the core mechanisms: encoding modalities into a shared latent space using transformers, fusing these representations through cross-attention, and employing contrastive learning for optimized model training.

The meetup concluded with a networking session and closing remarks thanking attendees from Abhilash and Prasad.

To stay updated with our future events and discussions:

• Follow us on Twitter/X - @bangpypers
• Join our Discord community: Discord Invite
• Follow our meetup schedules: Meetup Page
• Subscribe to our mailing list: Mailing List
• Follow us on Linkedin - @bangpypers