Internships at IBM in New York

The project will cover the integration of 2 IBM Research projects. The first (CADS) aims to find the best Machine Learning models given a certain dataset. The second project (Runway), stores and manages Machine Learning models and all its related data and links. The goal of the project is to keep track of CADS training and its results in the Runway system, such that the various CADS runs, models, and related data can be discovered, re-used, and served. On top of common database discovery operations that may use tags and keywords for discovery, this ML model management system also includes the extraction and representation of the various model performance metrics and model characterization. This facilitates model management from a ML point of view, for both the data scientist, as well as the scoring service/app developer.  Specific work items of this project include the design of the representation of ML model and CADS specifics, the implementation of such a respresention, the implementation of a CADS/Runway integration on either a VM or in IBM Bluemix, possible visualization aspects. The project can be finished with a demo and/or a paper submission that describes the system. Interested students may send a resume to : Horst Samulowitz (samulowitz@us.ibm.com).