Models that learn from data. What actually lands in the product — and what LLMs have replaced.
Category · Backend & Data
Models that learn from data.
Machine learning refers to algorithms that derive patterns from example data and apply them to new data — instead of every rule being hand-coded. Three basic types: supervised (with labels), unsupervised (without labels), reinforcement (with reward signal).
What of it lands in the product.
Day to day, mostly: classification (is this spam?), regression (what revenue next?), clustering (what customer segments exist?), anomaly detection (is anything going wrong?), recommendation (what next?). LLMs have reshuffled much of this — but not replaced it.
