Major Duties: • Develop software, typically in Python, to independently acquire data from disparate sources (databases, files, APIs, etc.) and combine them into appropriate training, validation and testing datasets • Analyze raw datasets using descriptive statistics, working directly with domain experts to understand the meaning of data fields • Build unit tests, data quality checks and data pipelines to ensure that algorithms use trusted data • Develop and maintain an understanding of many algorithms across supervised learning, unsupervised learning and time series analysis • Propose and develop machine learning ensemble methods that exhibit the best out-of-sample characteristics possible given the input dataset • Utilize expertise in machine learning algorithms to tune algorithms using available hyper-parameters and carefully select feature subsets • Discover biases or leakage in datasets and ensure that train/test splits reflect realistic expectations of real world performance • Run large scale (either in parallel and/or distributed) training and inference jobs on private or public cloud infrastructure • May present findings to internal and external customers using both data science language (F1 scores, regression error, statistical significance, etc.) as well as business domain specific language gained from experience analyzing the data in scope. Keras, TF, etc.), Azure cloud infrastructure preferred • Requires in-depth conceptual and practical knowledge in own job discipline and basic knowledge of related job disciplines • Applies best practices and how own area integrates with others • Explains difficult or sensitive information; works to build consensus Experience: • Computer Science degree (undergraduate or graduate level) and strong statistical background.