Data Scientist, Innovation Lab (37th Technologies)
I am a Data Scientist with Innovation Lab at 37th Technologies. I am an fiscally focused researcher first and an innovative analytics person next. In the past I
have worked in Insurance, Banking, Healthcare and Marketing domains.
My research is centered on Categorization of Human Sentiment on Social Media.
Sentiment Analysis in Unstructured Data - China
Ecole nationale de la statistique et de l’analyse de l’information - France
Kurukshetra University - India
Data Scientist at 37th Technologies
Apr.2017 - Present. India - Austria
Designed a context based chat-bot that would require each question to be designed as a tree by a set of rules and using the Partial Tree Kernel to find the similarities between the trees.
Data Scientist at Octopeek
May.2016 - Apr.2017.France
I was involved in Sentiment analysis of Twitter data. Use of text analytics approaches applied to the set of problems related to identifying and extracting subjective material in text sources. The entire project included testing on models like SVM, RandomForests, NB and MaxEnt. The accuracy achieved was 77% on the average of five times cross validation. Worked on French presidential elections of 2017 to gather and predict the voters sentiments with respect to candidates. Worked on a project for profiling the potential customers for an insurance company by using the Naive Bayes unsupervised models and making the use of INSEE data to gather the probabilities.
Business Intelligence - Associate Technology at Nagarro(Allgeier) - ERSTE Group
Jun.2014 - Jun.2015.India - Austria
Was involved in Fraud Detection in HLBA (House Loans Banking Application) utilizing Multilayer perceptron (MLP) neural network. Examined the application of Migration of Posting System and featured the key risk areas (data points) in the migration. Analyzed technical data from customers to highlight risks and recommend mitigation plans.
Claim Processing System - Software Engineer at UnitedHealth Group
May.2012 - Jun.2014India - United States
Developed a POC that empowered the business analysts across Business Units to monitor changes, frauds, the POC is a system based on unsupervised rule- based algorithm to scan health insurance claims in search of likely fraud. Designed and optimized a seamless data integration solution with which all applications interacting with Claim Processing System get up-to-date data all the time, resulting in savings to the tune of 100,000$. Was involved in developing key components of CPS across all business units, integrating business and partner systems encompassing 500+ common business services and 80m daily processing volume.