Jun 2018 – Apr 2020 · 1 yr 10 mos · Bengaluru, Karnataka

• Worked in market risk analytics team responsible to mine data driven insights from market activities
• Developed an algorithm to attribute changes in portfolio's value at risk to its multivariate risk factors using dimensionality reduction and correlation-driven clustering. Mentored a summer intern on the same.
• Automated trader marks validation process, by developing decision-tree based classifier. Identified key statistical features from historically classified marks in time series, to automatically action upon the newly observed marks.

May 2017 – Sep 2017 · 4 mos · New Delhi Area, India

• The aim of the project was to predict less-frequent high quality images from LANDSAT satellite using more frequent but less quality MODIS satellite images for better pest control.
• Implemented NASA's own data fusion algorithm iFSDAF in Python that would not require QGIS for co-registering images, first attempt of such kind.
• Also ended up improving the quality of the resulting images, tested over four similarity parameters. The project was in collaboration with agriculture-based clients, aimed to detect pest attacks via periodic imagery.

May 2016 – Jul 2016 · 2 mos · SIPCOT IT PARK, SIRUSERI, CHENNAI