Navigating the Complexity of Operationalizing ML Models

In this episode of Data Driven, our Andy Leonard and Frank La Vigne are joined by Chris McDermott, VP of Engineering at Wallaroo.AI. Together, they explore the challenges and advancements in the ever-evolving world of machine learning and artificial intelligence.

From the importance of ongoing care for machine learning models to the rise of edge computing and decentralized networks, they touch on the critical need for flexibility and data privacy. Chris shares his insights on the technical challenges of AI and ML adoption, as well as his unique career journey. They also discuss the evolution of technology and the potential future impact of these innovations.

Join us for a deep dive into the world of AI, technology, and the future of machine learning with Chris McDermott on this episode of Data Driven.

Show Notes

00:00 Exploring AI, data science, and data engineering.

06:20 Training and inferring are different stages.

08:12 Legacy AI doesn't require neural networks or GPUs.

12:09 Machine learning models require consistent care and monitoring.

15:10 MLOps merges skills, breaks down silos, collaborates.

16:47 Prefer MLOps to avoid namespace collision. DevOps parallels original Star Wars plot.

20:27 Internet-scale operations require automation and resilience.

24:13 Challenges of integrating AI into business processes.

28:03 New push for edge computing in technology industry.

32:05 Edge technology critical, discussed in government tech symposium.

34:50 Navigating from SendGrid to Twilio simplified processes.

36:15 First foray into data, growing knowledge.

39:33 Technology evolves, builds complexity over time.

44:41 Book recommendation: "Seeing Like a State" by James C. Scott discusses legibility and centralization of power in society.

46:28 Predictable tree farming fails due to ecosystem complexity.

Speaker Bio

Chris McDermott is a software engineer and entrepreneur who is passionate about creating products that make machine learning more accessible and manageable for users. His focus is on developing a platform that allows for easy deployment and management of machine learning models using any framework and on any architecture or hardware. He believes that current solutions in the market force users into a specific platform, and he aims to provide a more flexible and efficient alternative. With a strong belief in the potential of his product, Chris is dedicated to making machine learning more accessible and user-friendly for people across various industries.