Lessons Learned While Running ML Models in Harsh Environments

Explore the challenges and solutions of implementing machine learning models in critical financial environments through this insightful conference talk. Delve into the complexities of fighting organized crime in the financial sector, where billions of dollars are processed daily and downtime can have severe consequences. Learn about the various forms of financial fraud, including transaction fraud, stolen cards, anti-money laundering, and emerging scams. Discover the crucial balance between maintaining high detection rates and low false positives while ensuring system reliability, low latency, and high throughput. Gain valuable insights into data issues, scaling challenges, ethical considerations, system architecture, security concerns, compliance requirements, and business regulations. Understand the architectural tradeoffs and evolutions necessary for operating ML models in mission-critical environments. Benefit from the expertise of Pedro Bizarro, co-founder and Chief Science Officer of Feedzai, as he shares lessons learned from developing and implementing advanced fraud detection systems in the financial industry.

KDD2024 - Lessons learned while running ML models in harsh environments