In this episode, hosts Candace Gillhoolley and Frank La Vigne are joined by Neil Katz, Chief Product Officer at Valantor AI—a four-time Emmy winner whose unconventional journey spans from technology startups to award-winning journalism, and now to the forefront of enterprise AI innovation.
Neil shares his unique perspective on the evolution of AI, from the early days of digital design and machine learning, to building large-scale AI and document intelligence platforms for major organizations. The conversation explores the critical challenges of knowledge extraction, document comprehension, and securing sensitive data in today's era of sovereign AI. Together, they uncover the hidden complexities behind Retrieval-Augmented Generation (RAG), discuss the importance of hybrid search strategies, and reflect on where the field is heading as enterprise needs push the boundaries of what AI can do.
Whether you're a data professional, an AI enthusiast, or just curious about how language models are transforming how we understand information, you won’t want to miss this candid and thought-provoking deep dive into the future of AI and data integrity.
Links
Time Stamps
00:00 Early career in tech and journalism
04:01 Early consumer AI experiences
06:49 Early AI and machine learning developments
11:43 Anthropic's new findings on AI models
16:15 Data sovereignty in AI systems
19:27 Implementing open source AI models
22:49 Breaking down documents for models
25:45 Understanding the RAG system process
28:47 Challenges in AI data processing
31:22 Challenges in RAG with Insurance Data
36:35 Understanding and managing data security
39:35 Early days with OpenAI GPT
40:48 Explaining vector and similarity search
46:57 The evolution of computing models
47:29 Computing evolution to cloud and edge
