AutoML

AutoML, short for automated machine learning, uses software to automate repetitive parts of building a model. A system may prepare features, compare algorithms, tune hyperparameters, select architectures, allocate training resources, and rank candidates against validation metrics. This can help teams establish strong baselines and make machine learning accessible when specialist time is limited. Automation does not eliminate the need for judgment: people still define the problem, choose data, prevent leakage, select meaningful metrics, assess fairness, and decide whether the result is safe to deploy. AutoML can also consume substantial compute or overfit repeated validation tests, so its search process and final model should be documented and independently evaluated.