Anomaly Detection in Big Data

Apollo Platforms Group (now Wipro HOLMES), Wipro Ltd.

I joined Wipro’s Fraud Detection and Prevention and Anomaly Detection team, Apollo, as a part of the STAR batch, the fast track leadership development program consisting of engineering graduates from premier colleges of India.

  • I was responsible for the implementation and the execution of fraud detection rules which were run on the employee data to discover the potential anomalies for a wide variety of use cases, followed by building predictive models for fraud detection. The rules were executed on a single node Hadoop cluster (AWS EC2 instance) with the data being stored on AWS S3 and handled by Hive and MySQL.

  • I was involved in Apollo’s pilot for 3 potential clients, and responsible for the rule execution (on Hive), result validation and the final anomaly report generation (using JasperSoft).

Tools used: Apache Hive, MySQL, AWS EC2, AWS S3, JasperSoft