Last mile of AI: Continuously testing and delivery of models and software pipelines needs DevOps for AI
AI is being infused into software everywhere. Industry leaders are building and scaling trusted AI using prediction and optimization and deploying them through cloud native applications. As an organization increases the number of AI use cases and starts to deploy AI models with apps using DevOps, it's important to take a factory approach to AI. In particular, it is vital to get serious about testing and go beyond just looking at the models themselves. No testing means no trust in AI.
In this 60-minute live session, we will discuss the last mile of trusted AI delivery. Our featured guest, Diego Lo Giudici, VP & Principal Analyst of Forrester, will discuss methods and best practices of testing in AI systems. Diego will explore rigors of model testing, model complexity and multiplicity, variation in model performance over time and levels of uncertainty and autonomy. He also covers how to address the business risks of deploying AI with apps, testing frameworks, collaborations across DevOps, development and data science teams, lifecycle management and much more.
IBM will discuss the information architecture with a live demonstration to help unify processes, tools and talent to deploy AI at scale with Watson Studio on Cloud Pak for Data.
Key Takeaways:
- Discover how leading organizations are extending DevOps with ModelOps to increase yields of software development and drive digital transformation
- Learn potential areas of opportunities and challenges in testing and delivering AI inside an end-to-end lifecycle
- Explore best practices and methods for continuous integration and continuous delivery (CICD) of models and apps
- Learn how Watson Studio helps you get started and drive model and app delivery pipelines, including orchestration using Kubeflow.
Register now and get ready to lead the AI powered digital transformation by getting serious about AI testing and delivery!