Mes compétences :
Excel
Scala
Python
Spark
R
SAS Base
T-SQL
Entreprises
Veldos India Pvt Ltd
- Analyst
2013 - 2014To select the accounts to be sent to attorney and thus expedite the recovery process:
• Building/validating hypotheses which can impact propensity of debtors to pay the
amount
• Helped in the process of segmentation
• Developed the codes for extracting the top 5% accounts from the high score
segment which were supposed to be sent to attorney in the future
• Model maintenance
Gmid Associates
- Analyst
2012 - 2013To restrict the revenue leakage due to the claimants hiring attorney on BI(Bodily Injury) claims:
Roles and Responsibilities:-
•Helping model development team building/validating hypotheses which can impact propensity of claimant hiring attorney to represent their claims
• Developed the SAS codes for data extraction from various heterogeneous data sources and collating all information’s at claims level.
• Helped in building data Audit report and bi variate report which was further used by modelling team in outlier and missing value treatments and grouping/binning of variables under analysis
• Developed codes for various variable reduction techniques like correlation and multi collinearity and also helped in creating final dataset for modelling exercise
•Also developed and implemented the SAS codes to identify certain keywords from claim handler’s notes which were then used as predictor variables.
Gmid Associates
- Junior Analyst
2011 - 2012Claim Analytics:- Development of business scorecard to identify BI fraud cases for CFIT team
Roles and Responsibilities:
• Identification of key parameters which impact the propensity of fraudulent claims with discussion with claim handlers and business sponsors
• Designed and developed SAS codes for data extraction framework and collating all information from various data sources like policy data mart, claim data mart, claim handler notes and 3P data sources
• Bi Variate and hypotheses validation by data profiling against fraudulent claims
• Creating final reports where claim are sorted on no. of key indicators flagged and claims amount so that CFIT team can prioritize their claims investigations.
