The Microsoft News team is looking for a machine learning scientist who enjoy working on complex problems that push the boundaries of AI. We build one of the most popular news product that serves 56 different markets and hundreds of millions of active users. We mine and learn from petabytes scale data to deliver engaging experiences to our users that requires state of the art understanding of NLP, recommender systems, and Deep Learning. We solve problems such as: How do you analyze, understand, and encode user interest? How do you uncover the user's true intent from vague signals? Which key concepts best represent the content? How to encode a document into an embedding that can capture its essence? How to bring user understanding and document understanding together to construct personalized and engaging experience? We are a centralized team of scientists who attempt to solve these super challenging problems by leveraging our assets such as ample computation resources, internal collaborators
The ideal candidate has strong machine learning experience, moves fast to ship their techniques and likes to work on ambiguous problems that often need to be refined to collect a good quality data set for applying state of the art machine learning techniques.
2+ years of experience building practical machine learning models with applications to real-world problems. Experience with ML tools to build models and analyze data. (e.g. scikit-learn, keras, tensorflow) Knowledge of distributed computing systems (e.g. Spark, Hadoop, Hive, Cosmos) Demonstrated coding skills. Experience in Natural Language Processing and/or Recommender System is a plus. PhD in Computer Science or Statistics, or equivalent work experience