LLM prototypes are easy – production-ready LLM products are hard. Join us for a 90 minute hands-on workshop where you’ll set up everything you need for a reliable generative AI product at the AWS GenAI Loft. You’ll walk away with confidence in using tracing for debugging and optimization, setting up multiple types of evaluations, and optimizing for cost & latency, using the framework-agnostic W&B Weave LLMops tool.

Familiarity with Python will be helpful. No prior LLM product experience necessary.

What to expect:

Food, drink, peers building in the field, and a 90 minute hands-on workshop. You’ll end with a simple generative AI app that has solid observability via Weave and AWS Bedrock, evals that help you develop faster, and connections with AI builders at the exciting AWS GenAI Loft.

Topics covered:

• Logging input & output data in two (!!) lines of code
• Using traces to rapidly identify and fix problems
• Creating evaluations (programmatic, LLM-as-a-judge, and more) to score any aspect of your AI system
• Using evaluations to speed development and catch problems
• Using W&B Weave to improve speed, accuracy, and token efficiency
• Integrating Weave with Amazon Bedrock

Who should attend:

This workshop is for AI engineers, ML engineers, data scientists, and other AI practitioners looking to develop faster, cheaper, more reliable applications with LLMs. If you’re facing challenges with reliability, hallucinations, tone, bad structured data, model selection, regressions when making prompt changes, or more, this is for you.