Key takeaways

• How and when to introduce machine learning to your product (including how to validate MVPs quickly)
• How to collect and manage ever-evolving datasets
• How to measure, evaluate, and iterate on your progress
• How to maintain models as they get stale

Agile, Scrum, Lean… We have plenty of methodologies for traditional software development. But how do we tackle applied machine learning problems? Machine learning presents a new set of challenges for practitioners. How do we introduce ML in a product? Prototype and validate an MVP quickly? Collect and manage ever-evolving datasets? Measure, evaluate, and iterate on our progress? Maintain models as they get stale? These are some of the many questions professionals are facing in this field. This talk will cover methodologies, real-world best practices, and lessons learned for the applied ML life cycle.