Designing a machine learning tool, using text clustering, to investigate latency problems of report execution in the DFP reporting system on DoubleClick. Text clustering is used to find trends in report metadata. Python and scikit-learn are used for the project.

Worked on a project for catching fake accounts as part of Twitter’s anti-spam team. Used Java and Scala for coding the system as well as Scalding for mapreduce.

Consulted students if they had some issues while learning.

Made new problems for home works.