Databricks' vLLM Optimization for Cost-Effective LLM Inference - Ray Summit 2024
Discover strategies for optimizing vLLM to enhance LLM inference performance, reduce GPU idle time, and accelerate quantization, leading to cost-effective large-scale language model deployment.
Advanced Evaluation Techniques for AI Agents
Uncover advanced techniques for evaluating LLM-powered AI agents, addressing behavioral instability, comprehensive testing, and cascading failures. Gain insights from production experiences to implement robust evaluation pipelines.
Optimizing vLLM for Intel CPUs and XPUs - Ray Summit 2024
Dive into Intel's optimization strategies for vLLM on CPUs and XPUs, exploring technical advancements, performance data, and collaborative efforts to enhance GenAI inference efficiency.
Advancing Video Curation with Generative AI - NVIDIA's Approach
Explore NVIDIA's innovative approach to video comprehension in generative AI, focusing on data curation and multimodal video foundation models using Ray Data and NVIDIA DGX Cloud.
ByteDance's Platform for Reinforcement Learning from Human Feedback
Explore ByteDance's journey in building a video data processing pipeline using Ray's ecosystem for creating realistic video generation models from text instructions.
Ice Cream Animation in Blender 4.2 - Part 1
Master the art of creating a captivating ice cream animation in Blender 4.2 with this comprehensive tutorial, covering essential techniques and tools for realistic 3D modeling and animation.
Uber's Journey in Building Multi-Modal Foundation Models
Dive into Uber's approach to developing in-house multi-modal foundation models, optimizing earner onboarding and screening processes globally using Ray for distributed training and deployment.
Klaviyo's Journey to Robust Model Serving with Ray Serve
Explore Klaviyo's journey in building DART, a model serving platform using Ray Serve. Gain insights into architecture, deployment strategies, and practical advice for implementing Ray Serve in your own projects.
Faster Model Serving with Ray and Anyscale - Ray Summit 2024
Explore Ray Serve's capabilities for distributed model serving and deployment, focusing on Anyscale's solutions for large-scale AI models and the challenges of building AI applications in the era of generative AI.
Fine-Tuning Large Language Models at Scale - Workday's Approach
Dive into Workday's innovative approach to fine-tuning LLMs securely in a multi-tenant environment, addressing GPU scarcity and data access controls using PEFT techniques and KubeRay's autoscaling capabilities.
Using Promptable Image Analysis for Clinical Diagnostics - Ray Summit 2024
Explore the integration of LLMs and Vision Transformers for enhanced clinical image analysis. Discover how Ray technologies enable real-time processing of multimodal data, revolutionizing diagnostic workflows in healthcare.
Scaling Inference Deployments with NVIDIA Triton Inference Server and Ray Serve
Explore the integration of Ray Serve and NVIDIA Triton Inference Server for enhanced ML model performance and scalability. Learn to optimize inference deployments using advanced tools and a new Python API.
Handshake's Approach to Content Tagging with vLLM and Anyscale
Explore Handshake's journey in implementing vLLM-based content tagging using Anyscale, enhancing job feed performance and addressing growing LLM requirements through innovative internal tools and methodologies.
The State of vLLM - Advancements in LLM Inference and Serving
Explore the latest advancements in vLLM, the open-source LLM inference engine. Gain insights into its growing adoption, new features, performance improvements, and future roadmap for efficient LLM deployment.
Accelerated LLM Inference with Anyscale - Ray Summit 2024
Dive into Anyscale's latest LLM enterprise features and open-source contributions. Explore advancements in vLLM, including FP8 support and speculative decoding, doubling throughput and latency efficiency for improved inference performance.
Overcoming Distributed ML Challenges with Ray Train
Unlock efficient distributed ML pipelines with Ray Train. Explore solutions for setup, integration, debugging, fault tolerance, and data processing challenges in scalable machine learning workflows.
Building Scalable AI Agents with Ray
Explore Ragent, a framework for scalable AI agents built on Ray's distributed computing capabilities. Learn about its key features and integration with popular tools for developing production-ready agent applications.
Cloudflare's 150ms Global Cache Purge System - Deep Dive
Dive into Cloudflare's innovative global cache purge system, exploring its evolution from core-based lazy purge to coreless active purge, and understand the engineering behind its impressive 150ms performance.
Anyscale's Vision and Ray AI Compute Engine - Future of AI Scaling
Discover Anyscale's vision for AI scaling with Ray, the AI Compute Engine. Explore cutting-edge solutions tackling AI workload management challenges, from compiled graphs to advanced governance suites.
Foundation Models in Drug Discovery - Ray Summit 2024
Explore how foundation models are revolutionizing drug discovery across multiple biological domains. Gain insights into Recursion's large-scale approach using GPUs, robotics, and petabytes of data to transform pharmaceutical innovation.
A Brief History of Video Generation - Ray Summit 2024
Explore the rapid evolution of generative video with Runway's CTO, from recent breakthroughs to future possibilities. Gain insights into AI's transformative impact on audiovisual creativity and storytelling.