Beyond the Prompt: Evaluating, Testing, and Securing LLM Applications
Ever tweaked your LLM prompts or RAG pipeline and had zero idea whether it actually improved anything? In this NDC Copenhagen talk, Mete Atamel shows you how to measure real impact with evaluation tools like Vertex AI Evaluation, DeepEval and Promptfoo.
He also covers how to lock down your app against prompt injections and rogue outputs, diving into security frameworks like LLM Guard so your model sticks to the script and avoids any nasty surprises.
Watch on YouTube
Top comments (0)