M.S in Computer Science, New Jersey Institute of Technology
- Built an AI framework consisting of a Deep Convolutional Neural Network for identifying patterns in the historical datasets.
- Improved the analyst consensus from 52% to 74% for predicting the Lines of Influence in Stock Market using the Natural Language Processing framework and financial data.
- Extract meaningful topics out of a corpus of large text data or real-time streaming data. Reduced the manual human effort for hierarchical topic discovery from an average of 140 individual hours across 500,000 to 1 million tokens to less than 2 minutes.
- Led a team of data scientists and software engineers to source data from multiple data sources such as DBPedia, WikiData, NewsAPI, etc., and helped them perform ETL transformations/analytics on it.
- Improved the business value and efficiency for automatically validating the incoming text data from multiple sources and the benchmark was more than 95% of improved accuracy.
- Summarized huge amounts of real-time streaming multi-modal data with regards to the specialized topics that can change on a daily basis. (Approximately 9 billion tokens of text data were summarized over a period of 24 hours).
- Artificial Intelligence Researcher & Data Science Certified
Tata Consultancy Services