Rohit Kumar


About Me

July 2018 - now

  • Big Data Researcher, Eurecat

    • Working on multiple H2020 project in big data platform.
    • Contributing to Decode project
    • Big Data architecture design and data engineering

Aug 2014 - June 2018

  • PhD Student, ULB and UPC

    • Finshed 1 year research visit in UPC Barcelona Tech.
    • Funded by FNRS Scholarship

June 2014 - July 2014

  • Software Designer, RBS India Development Center

Nov 2007 - June 2014

  • Assistant Consultant, Tata Consultancy Services (TCS)

    • Technical Team lead (2011-2014) for online assessment platform offered by TCSiON: Digital Assessment
    • Researcher at TRDDC (2011). Working on data privacy.
    • Architect and Project Lead (2008-2011) for an Online Assessment platform development.
    • Teaching Assistant (2007-2009) at TCS iGnite.

Contact

rohit4phy AT gmail Dot com

CV

CV

Research

My research focuses on Graph Data stream mining. I work on developing algorithms for efficient graph stream processing which has application like Influence Maximization for viral marketing in social data stream. As part of my PhD, I have worked on supporting distributed graph stream processing on in-memory distributed graph processing systems like GraphX or Giraph. Currently, I am working on NoSQL systems and how to handle massive mobility data on NoSQL systesm.

Selected Publications

Conference

  • 2SCENT: An Efficient Algorithm for Enumerating All Simple Temporal Cycles. Rohit Kumar and Toon Calder. Proceedings of the VLDB Endowment VLDB 11 , August, 2018, Rio De Janerio, Brazil. [PAPER][Presentation][Code]

  • Activity-Driven Influence Maximization in Social Networks. Rohit Kumar, Muhammad Aamir Saleem, Toon Calders, Xike Xie and Torben Bach Pedersen. The European Conference on Machine Learning and Knowledge Discovery in Databases ECML/PKDD (Nectar Track) , September 18-22 , 2017, Skopje, Macedonia. [PAPER]

  • Cost Model for Pregel on GraphX. Rohit Kumar, Alberto Abello, and Toon Calders. 21st European Conference on Advances in Databases and Information Systems ADBIS , September 24-27 , 2017, Nicosia, Cyprus. [PAPER][Presentation]

  • Information Propagation in Interaction Networks. Rohit Kumar and Toon Calders. 20th International Conference on Extending Database Technology EDBT , March 21-24, 2017, Venice, Italy. [PAPER][Code][Presentation]

  • Location Influence in Location-based Social Networks. Muhammad Aamir Saleem, Rohit Kumar, Toon Calders, Xike Xie and Torben Bach Pedersen. Tenth ACM International WSDM Conference, 2017, Cambridge. [PAPER][Code][Poster]

  • Maintaining sliding-window neighborhood profiles in interaction networks. Rohit Kumar, Toon Calders, Aristides Gionis, and Nikolaj Tatti. Joint European Conference on Machine Learning and Knowledge Discovery in Databases (ECML/PKDD), 2015, Porto, Portugal. [PAPER][Presentation][Poster][Code]

Workshop/Demo

  • Cost Model Based Approach for Graph Partitioning in Spark GraphX. Rohit Kumar, Alberto Abello, and Toon Calders. Dutch Belgian Database Day 2017 (DBDBD) , 2017,Eindhoven, Netherlands. [Paper][Poster]

  • IMaxer: A Unified System for evaluating Influence Maximization Mechanisms in Location-based Social Networks. Muhammad Aamir Saleem, Rohit Kumar, Toon Calders, Xike Xie and Torben Bach Pedersen. International Conference on Information and Knowledge Management CIKM (DEMO) , November 6-10, 2017, Singapore. [PAPER][VIDEO]

  • Finding simple temporal cycles in an interaction network. Rohit Kumar, Toon Calders. The European Conference on Machine Learning and Knowledge Discovery in Databases ECML/PKDD (Workshop) , September 18-22 , 2017, Skopje, Macedonia. [PAPER][Presentation]

  • Time constrained Influence Maximization on temporal network(Poster presentation only). Rohit Kumar and Toon Calders. Spring Workshop on Mining and Learning 2016 (SMiLe) , 2016, Titisee, Germany. [Poster]

Patents

Software

IMaxer is a java based web tool. Its a system that unifies and combines different models and algorithms for experimenting and evaluating information propagation and influence maximization techniques in Location based social network. [CODE]

SDSLibrary is scala based stream data structure library which consist of implementation of HyperLogLog, CountMinSketch and Sliding HyperLogLog implementations. [CODE]

Academic Contributions

Teaching Proffesor

Course in Big data infrastructure for Master in Big Data Solutions, Barcelona Technology School, 2019-2020

Course in Big data systems for Master in Data Science Fondation, University of Barcelona, 2019-2020

Python for Data Science Course in Digital Vidya, India 2017-2018 (Online Live)

Teaching Assistant

Data warehouse for IT4BI Masters course in 2014-15 and 2015-16. ULB

Algorithms for TCS ignite trainees in 2008.

Java for TCS ignite trainees in 2008-09.

Reviwer

ICJAI 2020, DSAA 2019, TKDD 2018

External Reviwer

DASFAA 2017, MEDI 2016, DaWak 2016

Skills

Programming

Java, J2EE, Scala, C++(basics), MySQL, Haskel

HPC

Apache Spark, MongoDB, Oracle Coherence, Hadoop

Academic Links

Linkedin

Visitor Count

Flag Counter