There is nothing to show.
Designation:Data Analytics Consultant
Company:The Harlem Bee
Description:• Conducting analysis and predictive modeling of NYC public datasets in Python. • Providing classification, pattern recognition, forecasting based meaningful discoveries at requisite depth for multi-domain business development in Harlem and seed funding improvement by approx. 15%.
Designation:Data Specialist
Description:• Using Agile Methodology developed returned/exchanged orders application in Abinitio and Teradata overcoming 85% order misclassification and 10% data leakage. • Performed requirement analysis, designed modules, wrapper scripts using UNIX and SQL for BI report generation. • Conducted root cause analysis to resolve QA issues in multiple applications in coordination with Vodafone’s BI analysts. • Implemented ETL mapping, parallelism, data checks, fault tolerance techniques through Abinitio graphs and database objects (tables, views) in Teradata and Oracle DB ensuring overall EDW integrity.
Designation:Graduate Research Assistant
Company:Indiana University
Description:• Under the supervision of Prof. D. Williamson, implementing deep neural networks(DNNs) based speech separation and speaker recognition system using matlab, python, and tensorflow. • Preprocessing co-channel signals for noise reduction using complex ratio masking, then frame level log-spectral features as input and true speaker identities as training targets. • Also, using the ideal binary mask for frame level energy calculation of each speaker from the mixture cochleagram and to use their ratio for the soft labels in softmax output layer.