Eye On A.I. #216 Nenshad Bardoliwalla: Inside Vertex AI - The Ultimate AI Toolkit by Google Cloud
17 snips
Oct 30, 2024 Nenshad Bardoliwalla, Director of Product Management for Vertex AI at Google Cloud, dives into the transformative power of AI tools in business. He details the three layers of Vertex AI: the Model Garden, Model Builder, and Agent Builder, each enhancing model selection and optimization. Nenshad discusses the rising roles in AI model evaluation, stressing collaboration among tech professionals. He also highlights the importance of human oversight in AI design and compares the efficiency of Google’s Tensor Processing Units with traditional GPUs.
AI Snips
Chapters
Transcript
Episode notes
Vertex AI's Three Core Layers
- Vertex AI consists of three core layers: Model Garden, Model Builder, and Agent Builder.
- These layers provide model access, customization, and agent development with unique Google capabilities like web search grounding.
Choosing the Right Model
- Choosing an AI model depends on use case, latency, cost, and regional or task-specific needs.
- The Model Garden helps surface key info for users to narrow down model choices effectively.
Who Chooses AI Models?
- Experimentation with generative AI models can be done by software engineers and business analysts via prompting.
- Data scientists should perform formal evaluations to ensure the model fits enterprise needs.