Every week, new LLMs and techniques are released—Llama, Gemma, RAG, fine-tuning, and more—and it’s increasingly difficult to determine which approach will deliver the best results for your AI product. We all want to ship reliable products quickly, but testing every method while keeping up with rapid development cycles can be challenging.

In this talk, we’ll explore how you can leverage OpenFeature to perform A/B testing across different LLMs and AI techniques to identify what works best for your use case. Through a live demo, we’ll show you how to monitor key metrics like prompt reliability, model performance, and resource consumption, helping you make informed decisions on which approach to adopt—without compromising on speed or quality. This session is perfect for teams looking to experiment with different AI models and techniques in a controlled, scalable way.