May 2024 Summaries

Join us at Databricks Data+AI Summit to see the latest innovations in data and AI convergence, where you can learn from leading GenAI experts. Galileo will showcase its end-to-end capabilities for rapid evaluations, real-time monitoring, and GenAI protection, highlighting its capabilities at Booth #53 in the Expo Hall. The summit also features a session by CEO Vikram Chatterji on mitigating LLM hallucination risk through research-backed metrics on June 11th from 2:50 to 3:20 PM. Attendees can schedule 1-on-1 time with Galileo experts to explore how their team can build and evaluate GenAI apps faster.

We've developed Galileo Protect, a real-time GenAI firewall solution that intercepts hallucinations, prompt attacks, security threats, and more in real-time. This solution is designed to stop potential problems before they impact operations, providing customers with a 30% boost in efficiency and a dramatic reduction in response times. To use Galileo Protect, teams can easily configure a collection of rules and associated actions using the research-backed metrics or custom metrics of their own, monitor active stages and receive alerts when they are triggered, and refine and adjust the rules as needed. The solution comes packed with features designed to help enterprise teams quickly configure and manage a portfolio of GenAI safeguards and interceptions, ultimately helping build more secure, compliant, and efficient enterprise GenAI systems.

The AI landscape is rapidly expanding, with early winners emerging in various fields. RAG remains a dominant force in enterprise Large Language Model (LLM) systems. Recent research has consolidated existing knowledge on RAG and explored its evolution, suggesting new areas for innovation. AI-native enterprise apps have found success, with common traits among these winners providing a guide for generative AI companies. Researchers are also using AI to analyze disaster scenes and improve disaster response. Additionally, advancements in LLM pruning techniques and retriever methods can help build more effective RAG systems.

Galileo Protect is an advanced GenAI firewall that intercepts hallucinations, prompt attacks, security threats, and more in real-time. This technology aims to proactively prevent enterprise AI teams from being affected by hallucinations before they reach end users. The goal of Galileo Protect is to provide a robust defense mechanism against various types of attacks on AI systems. By detecting and mitigating potential threats, Galileo Protect helps ensure the integrity and reliability of AI-powered applications.