Title: Credit Card defaulter prediction.

Description: As part of risk analytics initiative project, We applied Machine learning techniques to address financial risk related problems. We developed a ML model using logistic regression to predict defaulter.  

Title: Sentiment Analysis using Text Mining

Description: Detected how the consumer is reacting to newly launch banking product, Performed sentiment analysis using overall conversation about them as well as product-related discussions. Deeper analysis of product-related discussions gives banks detailed insights. We developed model using Spacy

Responsibilities:

     Developed python code to apply text mining techniques using Spacy and Gensim libraries

     Executed the customer insight using Natural Language Processing, stemming and lemmatization, named entity recognition, text classification techniques

     Validated the models using statistical methods and interacted with data architects and clients to get approval

Title: Customer Credit Scoring using R and H2O.

Description: As part of HSBC Retail Credit Risk analytics, We applied Machine learning techniques to get credit scoring . We developed a ML model using GBM algorithm in R with H2O framework on a Microsoft Machine Learning Platform. The banking operations are using this model sectioning loans.

Responsibilities:

     Executed data pre-processing, interacting with data architects and client

     Developed model-ready data and machine learning model

     Validated each using statistical methods and tuned the parameters for better accuracy

Title: UAE customer income estimation.

Description: As part of HSBC Retail Credit Risk analytics for MENA region, We applied Machine learning techniques to estimate customer income. We developed a ML model using regression models R with H2O framework on a Microsoft Machine Learning Platform. The banking operations are using this model to target right customers for right products.

Responsibilities:

     Executed data pre-processing, interacting with data architects and client

     Developed model-ready data and machine learning model

     Validated each using statistical methods and tuned the parameters for better accuracy