
ML vs Software Engineering Complexity

Building Effective and Thriving Machine Learning Teams - David Tan & Dave Colls
Topic 3 of 9
Published Jan 27, 2025 · Clip 8:04–15:01 (6:57)
David Tan explains why ML systems are orders of magnitude more complex than typical software: data pipelines, compute scaling, monitoring model prediction quality beyond the four golden signals, and tying predictions to business impact. David Colls adds that ML features have less upfront certainty and a second vector of change from shifting data and world conditions, using COVID-era recommender systems as an example. The guests also clarify that ML here means supervised learning primarily, but also covers LLMs, reinforcement learning, simulation, and operations research.