Data Engineer Cum Scientist

Common Data Program - Sydney for Macquarie Group
A centralized program for bank to collate all the static, transactional and GL data into SQL Server so as to analyze and perform the analytics and build proper models to help bank to take better decisions.
KEY Roles:

• Created the high level design and plan from data sourcing to visualization for the CDP project in response to the business needs and requirements after discussion with different stakeholders in the bank.- in latest technologies using SQL Server, MSBI, AWS(Amazon Web Services)
• In conjunction with data owners, contributed towards development of data models and protocol to extract the data from RDBMS and dumping it to Hadoop/HDFC clusters using Sqoop.
• Wrote the complex hive queries to join and get the required set of data for analysis and to extract the key attributes for regression model building exercise.
• Used different types of data mining techniques to build a good model that could predict the active/inactive customers for the bank – the used machine learning techniques were – Decision Trees, Random forest, Neural Networks, K-means clustering etc using R-studio.
• Worked towards data consistency and integrity to make sure that the analysis meets the best of the standards to achieve better results.
• Collaborated with unit managers, end users and development teams and other stakeholders to integrate data mining results with existing systems.
• Engagement with senior stakeholders for on boarding new applications and technical design change and better solutions
• Working on Achieving and Performance tuning of workflows and batches
• Organizing the weekly meetings with project teams sitting in Sydney/Singapore/UK for bring everybody on same page.

Metrics and Analytics - RBS

Roles:
During different phases of project, played different roles 

- Modellers in R using advanced analytics/machine learning techniques –Logistic Regression, Decision trees, Random Forest, NN 
- as a database engineer with SQL/Hive on AWS plateform with Amazon,
- and a good presenter of results to the management. 
It helped bank to get profit around $50M in a quarter from the recommended suggestions.
- Used different types of data mining techniques to build a good model that could predict the active/inactive customers for the bank – the used machine learning techniques were – Decision Trees, Random forest, Neural Networks, K-means clustering etc using R-studio.
- Worked towards data consistency and integrity to make sure that the analysis meets the best of the standards to achieve better results.

SAPIENT - Singapore/India
Roles - 
Worked with different clients for Sapient for 7 years in data science and Analytics role - 

Risk Weighted Assets Analytics – Gurgaon/Singapore
Fraud Risk Analytics – Gurgaon/Singapore
Advanced Reports Analytics Engine Gurgaon 

- Data Scientist, Model Developer using R for regression and sometimes Neural Networks and other algorithms, wrote some of the reward winning models that accurately distinguish between good and bad buyers of card.
- Used different concept of Data engineering to gather the required attributes that could contribute to identity the fraud credit card applicant based on observation window.
- Used different sets to train and test datasets to build the model in R so as to build a correct model
- Used logistic regression / linear regression to identify the fraud applicants
- Created data definitions for new databases using SQL and sometimes changes to existing ones as needed for analysis
- Worked on streamlining the processes to change control and testing processes for modifications of analytics models.
- Worked on decision trees, random forest and logistic regression to build a better predictive model and validated the same using cross sampling the datasets.
- Extensive use of SQL to extract required data for business needs.
- To add a new report based on the user requirement and writing the stored procedures to retrieve the information from the data base primarily used SSRS.