Data Scientist at Z&A inc​

(February 2019 - present)

• Zettalabs, the product wing of Z&A inc., had two products-Zettasense and Zettainsights in its portfolio.

• I am working on development of various algorithms for solving record linkage or deduplication problem along with data standardization and normalization across various data sources and data formats.

• Apart from these two products, our team is also involved in game schedule optimization for NFL (USA). The scheduling with NFL constrains is a trillion possibilities problem and our team is developing algorithms for massive parallelization to generate and validate the game schedules. I am responsible for solving permutations and combinations math and validating the generated schedules using distance metrices.

Research Scholar at Anna University

(January 2013 to February 2019)

• Time series satellite imagery and socio-economic datasets were processed to identify the drivers of land use change.

• Developed two open-source packages ‘modifiedmk’ and ‘trendchange’ in R-language to perform non-parametric trend analysis on time-series data (More than 20k downloads from CRAN).

• Correlation and regression analysis, dimensionality reduction techniques were used to identify the associations and dissociations in the data.

• Future Land Use predictions maps were generated using Cellular Automata (CA), Markovian and Agent-Based Models (ABM).

Junior Research Fellow at ISRO

(March 2012 to November 2012)

• Worked as research follow in ISRO-GBP project- Monitoring land use dynamics and impacts on human dimensions at Indian River basins.

• Modelling the past and future changes in land use using supervised and unsupervised methods.

• Collected primary data and secondary data for identifying the impact of human dimensions on land use change.