Multimodal Engineering Night - SF Community Meetup

Multimodal Engineering Night is a meetup for engineers working on AI beyond text: images, 3D scenes, game state, and real product signals.

Understanding 3D Worlds Without Rendering Them

Harish Wajjala, Principal Machine Learning Engineer at Roblox, will explore how asset and position data can support scene-level embeddings for questions and classification, sidestepping the render-then-image pipeline. He will connect the approach to Roblox Economy use cases, including avatars and in-game transactions, and distinguish what is partly in production from what is still being explored.

Beating Video Games with Vector Search

John Kupchanko from Qdrant will walk through building an agent that uses vector search to play and win: embedding screen frames and game state, retrieving relevant memory, and making the engineering decisions that help an agent finish the game.

Community demos and networking

There are three five-minute demo slots for multimodal projects. Message the hosts through the registration page with a one-line description of what you would show.

Join builders working on computer vision, speech and audio, 3D and generative media, AI agents, search and relevance, embeddings, and vector databases. The evening includes food, drinks, and networking. Bill Raymond is one of the event's co-hosts.

The event description states that space is limited to 70. Registration requires host approval, and the exact address is shared after registration.