Data Science Manager - Credit Risk Modeling in Gurgaon, भारत

Key Responsibilities : Model Development & Research : - Credit Risk Modeling : Design and develop end-to-end credit risk models, including application scorecards, behavioral model development, and portfolio risk modeling.- Portfolio Analytics & Strategy : Monitor portfolio health across vintages, segments, and products through delinquency trends, roll-rate analysis, and early-warning indicators; translate these insights into risk-based pricing, underwriting policy, and credit strategy recommendations.- Supervised Machine Learning : Apply advanced supervised machine learning techniques alongside traditional statistical frameworks like Logistic Regression, Generalized Linear Models (GLM), and XGBoost.- Unsupervised Learning : Utilize unsupervised learning techniques like PCA (Principal Component Analysis) for dimensionality reduction and K-means customer clustering to identify risk segments, fraud vectors, and behavioral patterns.- Risk Metrics Estimation : Own the development of core risk parameters, including Probability of Default (PD), Loss Given Default (LGD), and Expected Credit Loss (ECL).- Performance Benchmarking : Benchmark and continuously improve model performance using appropriate evaluation metrics and experimentation https://jobeax.com/link/q5IOkHSeZ2N3p1Qj Handling & Feature Engineering : - Large-Scale Data Processing : Work with large-scale structured and unstructured datasets to build robust data pipelines, conduct exploratory data analysis, and engineer high-quality features.- Credit Bureau Integration : Leverage deep familiarity with Credit Bureau data to extract, preprocess, and incorporate bureau features into predictive workflows.- Data Quality Assurance : Ensure rigorous data preprocessing, augmentation, and validation to maximize model accuracy and https://jobeax.com/link/rPlV7TlOT7Khfhab Design & Optimisation : - Production Optimization : Collaborate with cross-functional teams to design and implement scalable algorithms optimised for production environments.- Model Explainability : Ensure risk algorithms conform to high standards of transparency, interpretability, and statistical https://jobeax.com/link/utjJ2ZzOUmZOMqwP Management : - Business Alignment : Partner closely with product, engineering, and business teams to align AI/ML solutions with organizational goals.- Scoping Requirements : Translate ambiguous business and portfolio problems into well-scoped data science problem statements with clear success https://jobeax.com/link/gPXmMmi05H8qXVo1 & Knowledge Sharing : - Team Guidance : Mentor junior data scientists and ML engineers, providing guidance on modelling approaches, code quality, and production https://jobeax.com/link/7CDpLslUuZoMb0Be Qualifications : - Bachelor's or Master's degree from a premier institution in Computer Science, Statistics, Mathematics, Data Science, or a related quantitative https://jobeax.com/link/nnF4EyWTLvU0XkJo : - 7 - 9 years of relevant experience in credit risk modelling and portfolio analytics, applied data science, machine learning engineering, or risk analytics specifically within the BFSI and lending domain.- Proven track record of developing credit risk scorecards, behavioral models, and regulatory frameworks (BASEL-2 / IFRS9 / ECL / PD / LGD).- Demonstrated experience implementing supervised machine learning techniques and unsupervised learning techniques (e.g., PCA, K-means customer clustering).- Strong understanding of statistical fundamentals and neural network fundamentals.- Generative AI : Experience or familiarity with Agentic AI frameworks and Retrieval-Augmented Generation (RAG) architectures.- Deep Learning : Hands-on experience applying deep learning techniques to financial services or credit risk use cases.- Advanced LLM Frameworks : Familiarity with prompt engineering, RLHF, and LLM evaluation frameworks.- Governance : Contributions to open-source ML projects, published research, or active participation in responsible AI and strict model governance practices within regulated industries. (ref:hirist.tech)