Making Open Models 10x Faster and Better for Modern Application Innovation

Learn how to accelerate open-source language models by 10x for modern application development in this conference talk from the AI Engineer World's Fair. Discover Fireworks AI's enterprise-grade deployment solutions that address the critical challenges of secure, low-latency, and cost-effective LLM serving for real-time generative AI applications. Explore the proprietary FireAttention technology that delivers 4x-15x faster performance compared to open-source alternatives, and understand how Fireworks' SaaS platform provides low-latency inference and high-quality fine-tuning across 100+ models including state-of-the-art LLMs, image/video/audio generation, embedding, and multimodality models. Examine the FireFunction model that integrates hundreds of models with API calling capabilities, and gain insights into software stack optimization for maximum performance across different hardware and deployment options. The presentation covers practical strategies for transitioning to AI-powered business applications through interactive experimentation and production-ready platforms built on PyTorch technologies, delivered by Fireworks AI's co-founder and CTO who is also a PyTorch core maintainer with extensive experience scaling PyTorch from research to production across Meta's AI use cases.