LLMOps is the set of practices for developing, deploying, monitoring, evaluating, and maintaining applications built around large language models.
LLMOps extends MLOps with concerns such as prompts, retrieval, context windows, guardrails, evaluation datasets, output monitoring, and cost management.