It’s essential for an AUV to interact well with its environment to have performant machine learning models and software. To help us, we are using Apache-Airflow to automate our machine learning pipeline and our data transformation pipelines (ETL) are fully automated. We are also using Labelbox’s distributed labeling features to label all the images we get from tests. (Tools and/or languages: Airflow, Labelbox, Google Cloud, Tensorflow)