Episode
959: Building Agents 101: Design Patterns, Evals and Optimization (with Sinan Ozdemir)
- Published
- Jan 20, 2026
- Duration seconds
- 3889
- Processing state
failed
Actions
POST https://stenobird.com/v1/public/podcasts/super-data-science/episodes/959-building-agents-101-design-patterns-evals-and-optimization-with-sinan-ozdemir/transcription-requests
Idempotently request low-priority transcript generation for this episode.GET https://stenobird.com/podcast/super-data-science/959-building-agents-101-design-patterns-evals-and-optimization-with-sinan-ozdemir.md
Read the agent-friendly Markdown representation of this episode resource.
Summary
AI entrepreneur and bestselling author Sinan Ozdemir speaks to Jon Krohn about the practical differences between agentic AI and AI workflows, why evaluating accuracy on its own won’t tell you enough about AI models, and more about his latest book Building Agentic AI. This episode is brought to you by the Dell, by Intel, by Fabi and by Cisco. Additional materials: www.superdatascience.com/959 Interested in sponsoring a SuperDataScience Podcast episode? Email [email protected] for sponsorship information. In this episode you will learn: (04:57) Exploring the differences between workflows and agents (17:03) How to work out parameter count for a given task (25:26) The best way to evaluate LLMs (33:12) How to run hybrid workflow + agentic projects effectively