A December 2017 interview with Dr. Debanjan Mahata, a 2015 graduate of the UA Little Rock Computer and Information Sciences PhD program. Dr. Mahata was recently hired as a Senior Machine Learning Engineer by Bloomberg Limited Partnership. Bloomberg LP is a privately held financial software, data, and media company with headquarters in New York city and owned by Bloomberg, Inc.
Question: How is Bloomberg using machine learning and what is your role?
Answer: Bloomberg has been using a lot of machine learning to enhance its products and to build cutting-edge solutions. The company has been very generous in supporting research initiatives in academia as well, through its Data Science Research Grant program. Please, find attached a flyer that gives more information about Bloomberg’s machine learning initiatives and recruiting.
I work as a Senior Machine Learning Engineer at Bloomberg, where I really enjoy my work. My role at Bloomberg allows me to work at the intersection of natural language processing (NLP), information retrieval, machine learning and software engineering. Entity extraction, entity resolution, disambiguation, keyword extraction, summarization, search and information retrieval from unstructured data are just some of the topics that interest our group. We not only research these challenging problems, but also build real-world solutions around them. The resulting applications ship as products in the Bloomberg Terminal, enabling our clients around the globe to make smarter, more informed decisions about their business and financial strategies.
I play an important part on the team that solves problems related to natural language processing from unstructured text. I am currently working on text summarization and classification. In this process, I get to work with very senior researchers who are famous in their fields, especially when it comes to formulating a research problem and discussing potential solutions. At the same time, when it comes to integrating a solution in the actual product, I get to collaborate with very talented software engineers, who follow a careful process of rigorous validation.
Bloomberg offers a very stimulating work environment for machine learning engineers like me. They not only provide a great experience for learning the best practices of software engineering, but also encourage its engineers to participate in machine learning conferences and workshops. Recently, they provided a sponsorship of MR2AMC: Multimodal Representation, Retrieval, and Analysis of Multimedia Content, a workshop that I am organizing along with some of my research collaborators in conjunction with the IEEE’s International Conference on Multimedia Information Processing and Retrieval.
Question: What was your previous professional experience?
Answer: Before joining Bloomberg, I was a Senior Research Associate at Infosys in Palo Alto, California. My primary role there was to do research and development work in natural language processing for the Nia Platform team. Some of the things I worked on included applying deep learning techniques to tasks such as phrase embedding, keyword extraction, text classification, text summarization and semantic similarities between text documents. Some of the research I did there produced state-of-the-art results and led to the invention of some novel techniques related to automated keyword extraction from text documents using deep learning methods. This led to the filing of patents and publication of research.
Question: How did your education at UA Little Rock helped prepare you for this role?
Answer: Whatever I work on today and have been doing for the past two years are related to my education and training that I received as a Ph.D. student at the University of Arkansas at Little Rock. Courses such as Data Mining, Entity Resolution, Social Media Mining, Research Methodologies and Information Quality Tools have immensely helped me dive deep into the realm of machine learning research, in addition to searching for answers to difficult problems.
Some of the things taught in the Entity Resolution course are very unique and directly relate to some of the work I presently do at Bloomberg. My training as a research scholar under Dr. Talburt gave me the necessary skills to think independently and to attack unsolved problems in creative ways. Feel free to look at my doctoral dissertation, “A Framework for Collecting, Extracting and Managing Event Identity Information from Textual Content in Social Media.”
As an affiliate graduate faculty at UALR, I would love to motivate and guide students to learn and work on problems related to machine learning. I would also welcome collaborating on research with students and fellow faculty members.
The Computer and Information Sciences PhD program is offered through the George W. Donaghey College of Engineering and Information Technology at the University of Arkansas at Little Rock. The information science and information quality tracks of the program are now available through the online eLearning Campus.