Building AI Applications with LLMs

In this 39-minute conference talk from Conf42 Python 2025, learn practical strategies for developing robust AI applications with Large Language Models. Discover how to overcome common challenges like LLM inconsistencies through proper implementation of retries and timeouts. Master techniques for handling structured outputs, implementing streaming responses for improved user experience, and managing long-running tasks with background jobs. Explore methods for evaluating prompt effectiveness and gain insights into various foundational models. The presentation provides a comprehensive walkthrough of best practices for building reliable AI applications, covering everything from initial development considerations to advanced implementation techniques.