As algorithms increasingly make decisions about hiring, lending, criminal justice, and healthcare, the question of fairness becomes urgent. This article introduces the key concepts of algorithmic fairness: different mathematical definitions of fairness (demographic parity, equalized odds, individual fairness), why they are often mutually incompatible, and the sources of bias in training data and model design. It provides a practical framework for fairness audits, bias mitigation techniques, and the organizational processes needed to embed fairness considerations into the ML development lifecycle.