LiFT, a Scala/Spark library from LinkedIn, provides tools to easily measure biases in training data and fairness metrics for ML model performance. Specifically, LiFT uses Spark to load input data into in-memory data structures and calculate predefined or custom metrics (specified through UDFs) in a distributed manner. Predefined training data metrics include those that measure distance/divergence from a given reference distribution (e.g. skews), metrics that compute distance/divergence between segments of the observed distribution (e.g. demographic parity), and aggregate metrics to evaluate higher-level notions of inequality(e.g. generalized entropy index). The model performance metrics additionally includes statistical tests for fairness and a permutation testing framework.