FOUNDATIONS • 36 / 397
Start before LLMs. Understand where AI, ML and neural networks fit.
Underfitting
Underfitting happens when a model has not learned enough of the real pattern to perform well even on familiar data.
Think of it like
Think of Underfitting as one layer in a very large toolbox: you only need to understand what job that layer performs.
Real life
A phone spam filter, recommendation engine, fraud detector or voice assistant may use Underfitting somewhere inside.
SRE lens
A weak anomaly model may miss obvious failures across both training and validation data.
Remember thisUnderfitting happens when a model has not learned enough of the real pattern to perform well even on familiar data.