Introduced on day 1
Inference (statistical)
Drawing conclusions about a population or a data-generating process from a sample, with a statement of uncertainty: which relationships exist, how strong they are, how sure you can be. In ML engineering, "inference" instead means running a trained model.
In detail
Statistical inference asks what the data say about the world beyond the sample: is the no-show rate on Friday evening flights really higher than on Tuesday mornings, or could the gap be noise? How strongly is no-show associated with fare class, holding route fixed? Answers come with intervals or tests, not just point values.
Inference about an association is not a causal claim. Suppose a regression on historical flights shows that no-show rates differ by fare class. The reason may be who buys each fare (for example, travellers who need flexibility), not the fare itself, so changing someone's fare need not change their behaviour.
Watch the word: in ML engineering, "inference" means serving a model to produce predictions ("inference latency", "batch inference"). In a meeting, check which meaning is in play before you answer.