Episode 17: AI Cloud Essentials Podcast
As agentic AI moves from experimentation into real-world applications, reliability, evaluation, and observability are becoming essential. Organizations need tools and infrastructure that help them understand agent behavior, optimize performance, and confidently deploy autonomous systems across complex environments.
In Episode 17 of AI Cloud Essentials Podcast, host Ritu Jyoti sits down with Uma, Nico, and Brandon to explore what it takes to build and scale production-ready AI agents. The conversation examines the challenges of long-running agent workflows, specialized use cases, and the growing importance of continuous evaluation and monitoring.
Learn how tools like Weights & Biases Weave support prompt engineering, agent evaluation, and observability, while CoreWeave provides the infrastructure needed to run and optimize AI workloads. The panel also discusses how model training, reinforcement learning, MCP, and emerging agent development tools are shaping the next generation of autonomous AI systems.
What you’ll learn
- Why longer agent workflows introduce new reliability and evaluation challenges
- How online evaluation and offline benchmarking improve agent performance
- Why observability and monitoring are essential for production agentic AI
- How Weights & Biases Weave supports agent development and optimization
- How model training and reinforcement learning improve agent behavior
- Why MCP and coding agents are becoming important parts of AI development workflows
Guests: Brandon Kates, Forward Deployed Engineer, Weights & Biases by CoreWeave
Uma Krishnaswamy, Senior Account Solutions Architect, Weights & Biases by CoreWeave
Nicolas Remerscheid, Product Manager, Weights & Biases by CoreWeave
