Split
Train
Evaluate
A better score can hide a worse experiment
Machine Learning2 min readDraft
Why the train/test boundary matters before scaling, feature selection or model tuning.
Read noteSoftware, AI &
the data between.
Practical notes on building applications, working with models and asking better questions of data.
Machine Learning
1 note across software, AI and data
Editorial drafts for review. Examples are illustrative and do not describe employer systems or project results.