LLMOps: Test-Driven Development for Large Language Model Applications

Practitioners should build LM Programs comprised of several composable calls to LLMs which can be rigorously tested, audited, and optimized. This talk covers the emerging engineering discipline of production machine learning and LLM applications.