DataHour: Enterprise Challenges with Generative AI and LLMs

The talk will cover three critical aspects of GenAI and LLM applications in an enterprise context.

1. Performance and Scalability

• Hands-on: Evaluating response time and throughput of services.
• Design choices that extend beyond local proof of concept to create an enterprise-level scalable service.

2. Security

• Hands-on: Detecting malicious prompts using LLM.
• An overview of security considerations in enterprise environments and potential strategies for addressing them.

3. Quality and Consistency

• Hands-on: Evaluating RAG pipelines
• Best practices to ensure the quality and consistency of services amidst design changes.
  
  

Sagar Patel brings over seven years of experience in designing ML and AI solutions across various domains. He boasts extensive expertise in developing text analytics platforms and services utilizing NLP. Currently pioneering Visa’s internal ChatGPT platform, his work focuses on adapting the latest developments in GenAI while upholding the highest levels of security. Passionate about teaching, he has delivered over 1,200 hours of training in ML and AI to both classroom and corporate audiences.