Avinashkumar Vellore Loganathan

• Led the AML Rule Uplift Program, optimizing financial crime detection by refining monthly query rules and manual reporting processes.
• Applied sophisticated statistical methods to create customer segmentation models, categorizing customers by Line of Business, products/services, and transactional behavior.
• Developed and deployed innovative rules based on segmentation insights, enhancing process efficiency and reducing false positives in financial crime detection.
• Partnered with diverse teams to set segment-specific thresholds, aligning monitoring efforts with regulatory compliance.
• Performed continuous Below The Line (BTL) testing to validate threshold efficacy and maintain compliance standards.
• Provided technical leadership in statistical analysis and data- driven strategies, elevating team proficiency in compliance practices.

• Led the development and implementation of the Financial Crime Client Risk Rating (FCCR) model, a critical initiative aimed at classifying customers into four distinct risk buckets based on financial crime risk factors
• This model significantly influenced KYC review frequency and monthly alert generation processes
• Orchestrated the entire lifecycle of the FCCR model, including data collection, cleaning, feature binning, sample preparation, model development, and ongoing monitoring
• The model was constructed leveraging geographies, client characteristics, reputation, and transactional activities
• Acted as a liaison with the Model Risk Management team, providing thorough model validation, hypothesis selection, and recommendations
• Led a team of four members throughout the project lifecycle, ensuring high-quality deliverables and adherence to regulatory standards
• Achieved an estimated cost saving of 800K Euros per year through the successful implementation of the FCCR model
• Developed automated Python scripts to seamlessly integrate the model's output into the production system, reducing manual effort by two weeks every quarter with 100% accuracy
• Spearheaded the development of False Positive Reduction models, leveraging typology analysis and SAR data from the past two years
• Designed a classification score model to predict the likelihood of an alert resulting in a SAR filing, utilizing KYC and transactional features
• Collaborated with the Model Risk Management team to validate the model's effectiveness
• Executed a project focused on the AFMO region, addressing missed transactions by the CBS system in AFMO regions for Suspicious Activity Report (SAR) activity
• Utilized three years of transaction data and applied the three sigma methodology to identify and flag suspicious transactions
• Filed SARs based on AML activities in Tunisia, Ivory Coast, and Senegal, conducting all statistical analysis using Python
• Exceptional project management skills, including team leadership and stakeholder communication

• Served as the Senior Data Analyst for the Home Buying project, with a primary focus on understanding the underlying data models within the Home Buying area of the Australian bank
• Collaborated closely with functional teams to gather and analyze data required for various analysis purposes
• Utilized SQL extensively for data extraction, transformation, and loading processes, ensuring the availability of accurate and relevant data for analysis
• Leveraged Python for data visualization purposes, creating insightful visualizations to communicate findings effectively to stakeholders and support decision-making processes

• Novu Responsiveness Model - Developed a predictive modeling using Random Forest by collecting responsiveness of the people and based on history of medical claims, appeals and grievances factors and demographics data.
• Urinary Incontinence Issue Prediction – Developed a simple logistic regression model for predicting urinary incontinence issue for the people age greater than 60. Major challenge was to identify the secondary disease diagnosis
• Propensity to reach Model – Developed a call responsive predictive model using Euclidean Distance approach by considering Clinical and Demographics information.
• Member retention Model – Developed a recommender system that will identify a potential group of customers based on Compliance(Scoring), Customer satisfaction(Sentimental Analysis). Used NLTK for sentimental analysis by taking history of calls a customer made

• Designing scripts in Teradata across 3 layers and developing the reports in BO as per the user requirement.
• Involved in requirements gathering from Europe business Architects and building the system.
• Developing the Automation jobs as a part of enhancements which helped the team to carry out the day to day support activities with ease.
• Experience in requirement gathering from Clients and Business Users
• Developed the Automation jobs in Unix Shell Scripting
• Provided Automated reports from source transaction systems using Tableau

• Developed the DataStage ETL jobs as a part of enhancements which includes extraction ,business validations ,transformations and loading into the warehouse
• Performance tunings in both DataStage jobs as well as in SQL.
• Involved in analyzing and fixing the data level issues.
• Coordinate daily status calls with client, SME’s and provide status about the daily batch flow apart from regular incidents we receive.
• Being an application SME, my role is fathoming the nuances of the batch flow and ought to
  provide an expert solution in order to improve the performance of batch.
• Automated the Outage process in Walmart which consumed more manual effort.
• Designed Warning extractor script from all DataStage jobs which saved 3 months of manual
  effort and gained Customer appreciation for delivering the project in a quick time.