Job Description

Job description

Cross-Functional Collaboration: Coordinates with Data Engineers to validate structural data, and aligns with project/account managers across disciplines.
Exploratory Data Analysis (EDA): Queries large-scale banking data assets to assess data quality, distributions, and availability before modeling.
Feature Engineering: Utilizes Python to transform pre-aggregated datasets into features using methodologies like one-hot encoding, frequencies, and lag variables.
DS Project Architecture and Design: Architects, builds, and operationalizes advanced predictive models for financial sector clients within Databricks.
Backtesting & Simulation: Executes a rigorous backtesting and simulation that evaluates the performance of existing detection rules against historical and synthetic datasets to quantify their precision and recall.
Fraud Detection & Optimization: Develops real-time fraud detection pipelines and actively performs false positive optimization.
Collections Analytics: Performs complex collections analytics including roll-rates analysis, vintage curves analysis, and recovery modeling / forecasting.
Leadership & Mentoring: Guides and trains direct reports on project-specific tasks and overall career goals.
Strategy: Participates in assessing business challenges and setting company direction/initiatives for analytics.
Portfolio Management: Oversees a portfolio of data science projects and guides technical teams as business needs require.
Project Documentations
Models: Robust machine learning models for Banking use cases (e.g., fraud detection, risk scoring, customer analytics).
Frameworks: Prevention strategy & policy frameworks.
Simulations & Evaluations: Rigorous backtesting reports quantifying precision and recall.
Model-ready Feature Sets: Engineered using standard Python libraries.
Life Cycle Operationalization & Management
Project Architecture and Design

REQUIREMENTS

A degree in Computer Science, Statistics, Mathematics and related fields.
Has 5 years or more related experience, with a strong, proven track record in building machine learning models, specifically focusing on Fraud Management (Card Fraud, Application Fraud, ATO) and Collections Analytics (roll-rates, vintage curves, recovery modeling).
Working knowledge on any programming language that can be used for data science and analytics like Python, R, Java, and the likes. For this engagement, strong proficiency in Python and SQL is highly required.
Deep, hands-on experience architecting, engineering features, and training models natively within Databricks, including life cycle operationalization.
Knowledgeable in CRISP DM or similar methodologies.