How would you optimize a machine learning pipeline that is processing large volumes of time-series telemetry data for Airbus Helicopters?
