Experience

Vission Technology Inc.
• Canada
• Information Technology & Services
• 1 - 100 Employee
• Technical Consultant and HR

  • 2017 - Present


Amazon

• Bangladesh
• Advertising Services
• 1 - 100 Employee

• Research Scientist

  • 2015 - 2016

  Used statistical and machine learning techniques to create or optimize scalable solutions for business problems; Designed, developed highly innovative models for predictive learning • Developed production models with AML (AWS Machine Learning) tool to improve matching rate, which is redefined based on analysis and business logic, using a great technique to justify usage of surrogate variables for easy implementation in production. Used statistical and machine learning techniques to create or optimize scalable solutions for business problems; Designed, developed highly innovative models for predictive learning • Developed production models with AML (AWS Machine Learning) tool to improve matching rate, which is redefined based on analysis and business logic, using a great technique to justify usage of surrogate variables for easy implementation in production.



M&T Bank

• United States
• Financial Services
• 700 & Above Employee

• Quantitative Risk Modeller II, AVP

  • 2013 - 2015

  Worked on CCAR (Comprehensive Capital Analysis Review) reports for Business Banking. Ran CART for hot spot analysis for automation of approval of Business Banking Built general behavior models for credit card account management with new insights. Worked on CCAR (Comprehensive Capital Analysis Review) reports for Business Banking. Ran CART for hot spot analysis for automation of approval of Business Banking Built general behavior models for credit card account management with new insights.



Capital One

• Egypt
• Leasing Non-residential Real Estate

• Senior Statistician

  • 2011 - 2013

  Conducting R&D on building a revenue model for acquisition. Involved in building a flag-ship risk model: proposing and justifying conversion of model scores to FICO like scores for standardizition, picking samples and simulating TDR (Turn Down Reason). Retro scored risk and revenue models, running regression models for calibration with missing variables resulting in great feedback from customers and stakeholders. Documented use of statistical models for a business project in MGS (Model Governance System), proposed a monitoring plan to cover a blind spot. Moved forward successfully a project by saving significant budget on programming outsourcing. Built valuation models to predict interest of auto loan product with better prediction. Performed one time validation on various models for HSBC Credit Card (purchased by COF) with excellent insights revealed. Provided ad hoc support with significant achievements and savings on resources Show less