Kunal Jha

Developed a ranking algorithm which is 7 times faster than PageRank using Spark Developed web services for all of the tools using Spring with Tomcat Server Built a rule mining tool using PREGEL on Spark’s GraphX API Built a data source enriching tool where each resource in the KG was enhanced with corresponding labels found in NL Prepared the design and unit and Integration testing and documentation Developed tool to correct human created gold standards for better evaluation of annotation tools

Fixed bugs and Optimised the working of the lookup tool for DBpedia upto 2 times the original code. Extended the querying and indexing mechanism to three languages and made plugins for further language. Packaged the tool with Swagger API and enhanced the setup with Docker that allowed the tool to be setup. Developed a prototype for a new ranking algorithm which was 2 times faster than PageRank for RDF data on single machine.

Developed a document similarity tool which is the core idea of the startup applying Smith Waterman algorithm. The tool can take one’s idea description and match it with all the documents that directly or indirectly discuss the similar concepts. The work used tools like MALLET and trained LDA models to apply the algorithm to get word, sentence and document level similarity.

Developed a decision tree model based backend logic for a student performance analysis app. Involved in testing the prototypes and planning the business model for the app

Intelligent Systems