In our podcast, we delve into the complex technical and societal aspects of the development and use of Generative AI. Through long-form pieces, roundtables, and interviews, we aim to bring you up to date on this fast-moving field, as well as discuss longer-term perspectives.
In this episode of the GenAI Days Podcast, we explore the challenges of testing GenAI applications with non-deterministic language models. Learn how to define test cases, generate prompts, and evaluate LLM responses using embedded snapshots for enhanced reliability. Fabien Zucchet, a Serverless Cloud software engineer at Aleios, shares his experience building GenAI applications and contributing to the development of testing methodologies for non-deterministic LLMs.
In this episode, we explore the risks associated with GenerativeAI (GenAI) and how companies can manage them effectively. We discuss the supply chain of GenAI applications, including IP and data security concerns, potential malicious use, and implicit bias.