Bloomberg’s core product, the Terminal, is a must-have for the most influential people in finance. In addition to being the second largest producer of news in the world, Bloomberg ingests more than 1.5 million news stories per day from more than 120,000 different sources to help our clients stay in the know. This data would be unmanageable without our help. News stories move markets. We build machines that understand them.

With the ever-increasing amount of data being generated, our clients need ever more sophisticated ways of extracting signal from noise. The Pattern Recognition team in the Machine Learning group is responsible for building a wide variety of analytics and recommendation systems to surface insights (e.g., identify market moving news stories, create new events that are relevant to users, provide personalized recommendations of events to users, etc.) to increase our client's productivity.

We are looking for research scientists with experience running a machine learning project from inception to completion, and specifically with a background and interest in one or more of the following areas:

• State of the art deep learning (e.g., hyper-parameter tuning, sequence-to-sequence methods, word/entity embeddings, CNNs, etc.)
• Model evaluation
• Feature engineering
• Big data experience
• Familiarly with time series analytics

We'll trust you to:

• Drive projects as the principle point-of-contact, with the ability to determine and elaborate on suitable metrics as well as methods
• Write, test, and maintain production-quality code (mainly C++)
• Publish papers and attend conferences in leading venues representing Bloomberg

You'll need to have:- A graduate degree in a quantitative field (PhD preferred but not required)

• Recent experience with C++ or Java
• Strong computer science fundamentals (algorithms, data structures)
• A track record of relevant research demonstrated by publications in NAACL, ACL, CVPR, ECCV, ICML, NIPS, ICLR, IJCAI, SIGIR, AAAI, KDD, or WWW
• Solid background in machine learning and/or statistics, publication history
• Experience with a deep learning framework (e.g., TensorFlow, PyTorch, Keras, CNTK, etc.) is helpful
• Experience with NLP is also very useful