​❗ This event is way over capacity, and we will have to turn folks away once we reach the cap because of safety concerns. Please do arrive early if you are truly interested in it. Please kindly cancel the registration if you are unsure about it.

​❤️ Thank you all for the enthusiasm and support.

​Join experts from Anyscale/Ray, LanceDB, and Character AI for an in-person deep dive into Multimodal AI on May 14th. Gain invaluable insights from pioneers shaping this cutting-edge field. Don't miss this chance to expand your knowledge, network, and stay ahead in Multimodal AI. Secure your spot for this exclusive gathering of Multimodal AI leaders and enthusiasts.

​Please find detail of the event below👇

​Event Details

​Date: May 14th, 5PM PT

​Location: Anyscale SF office. Space is limited, please do register to be allowed in.

​Agenda：

​5pm - 5:30pm Door Opening

​5:30pm - 6pm

​Scalable LLM Offline Batch Inference with Ray Data

​Cheng Su, Manager of Data Team, Anyscale 

​Zhe Zhang, Head of Open Source/Ray, Anyscale

​LLM offline batch inference with multi-modality is gaining attention and growing quickly. Ray community has developed Ray Data, an open source distributed ML data processing library (https://docs.ray.io/en/latest/data/data.html). Ray data takes unstructured data as first-class citizens, runs at scale on GPU and CPU. In this talk, we will introduce the basic concept and API of Ray Data, dive deep into streaming execution under the hood, and discuss targeted use cases including LLM offline batch inference, embedding generation, and ingestion to vector databases

​6pm - 6:30pm

​Empowering Multimodal AI from Desktop to Enterprise Scale with LanceDB-Ray

​Lei Xu, Cofounder CTO, LanceDB

​Jasmine Wang, Head of Ecosystem Engagement, LanceDB

​Enterprises in the GenAI sector have long sought a seamless solution for managing and extracting value from vast amounts of video and audio data. With the LanceDB-Ray integration, that solution is now within reach.LanceDB delivers unparalleled value with its zero-copy schema evolution capability and robust support for large blobs, simplifying multimodal data analytics to the level of SQL. With its advanced indexing and rapid random access, coupled with distributed training, the combined LanceDB-Ray solution streamlines high-performance ML training, delivering lightning-fast results with selective filters.Join us for an in-depth exploration of this partnership designed to unlock the full potential of multimodal data.

​6:30pm - 7pm

​How Character AI scale out the model research velocity and optimize inference batch

​Rong Rong, Member of Technical Staff, Character AI

​In this talk, we will cover the following on going work at Character AI:
- Scaling out research velocity with ray/vllm serving
- Batch inference optimization
- Ray-Lance use case at Character.AI

​7pm - 8:00pm

​Social networking