AI/ML Operationalization Anti-Patterns
AI/ML Operationalization Anti-Patterns

Abstract: 

Is your enterprise struggling to operationalize AI/ML workloads? You're not alone. In this session we'll go through the most common anti-patterns global organizations are struggling with. We'll cut through the hype and buzz words while using real-world customer examples of these anti-patterns and then discuss approaches that provide scale, agility, and efficiency Is your enterprise struggling to operationalize AI/ML workloads? You're not alone. In this session we'll go through the most common anti-patterns global organizations are struggling with. We'll cut through the hype and buzz words while using real-world customer examples of these anti-patterns and then discuss approaches that provide scale, agility, and efficiency for these AI/ML workloads.

We’ll first discuss anti-patterns from traditional big data deployments including data governance worst and best practices. Then, we’ll cover the most common deployment anti-patterns for big data environments and contrast those with a better way to scale in a cloud-like manner. But, the rush to the cloud is another common anti-pattern that has enterprises struggling to realize the value that the hyper-scaler providers can bring. The question every enterprise must face is where do we choose to take technical debt when using the cloud and how do we prevent vendor lock-in while not over-engineering the process so much that agility and productivity are crippled.

The final anti-pattern we’ll cover is the organizational approach to data science. Many organizations are struggling to reorganize to support this new discipline and are putting expectations on these highly skilled data scientists that cripple their productivity. Data science is a team sport, so well discuss how organizations should structure their teams and approach AI/ML projects from idea inception to development to operationalization to managing and monitoring and finally refactoring and refreshing their analytical models.

We’ll close the presentation with a pattern to solve these challenges and provide examples of world-class organizations that have implemented the pattern to provide enterprise scale, agility, and ultimately organizational transformation.

Bio: 

Matt Maccaux has been working with customers across many industries for the past 20 years at some of the biggest technology companies in the world. For the past 8 years, Matt has focused on the big data and analytics space, helping customers define and implement enterprise-wide programs to accelerate their time to market using advanced analytics. In his current role as the Global Field CTO for HPE's Enterprise Software group, Matt is working with executives to develop roadmaps and strategies for their next generation analytics using AI/ML/DL and providing those capabilities as-a-Service to the enterprise.