Senior Data Scientist with 7 years of experience delivering machine learning, predictive analytics, statistical modeling and AI solutions across BFSI, financial analytics and enterprise knowledge management. Experienced …
Senior Data Scientist with 7 years of experience delivering machine learning, predictive analytics, statistical modeling and AI solutions across BFSI, financial analytics and enterprise knowledge management. Experienced in credit risk modeling, collections optimization, climate risk analytics and Generative AI/RAG solutions. Strong hands-on expertise in Python, SQL, Scikit-learn, XGBoost, Random Forest, Logistic Regression, LangChain, RAG, LLMs, vector search, AWS SageMaker and Snowflake. Proven ability to translate complex data and business problems into scalable, interpretable and decision-ready analytical solutions.
<p>Developed a Climate Risk Assessment Model using Environmentally Extended Input-Output (EEIO) datasets to quantify sectoral and geographic climate-risk exposure, enabling data-driven portfolio risk assessment and strategic decision-making.<br>
Integrated Scope 1, Scope 2, and Scope 3 emissions to assess carbon intensity and supply-chain dependencies, improving visibility into direct, indirect, and value-chain climate exposures.<br>
Built sectoral and geographic risk segmentation frameworks using emissions-weighted exposure, enabling prioritization of high-risk industries and targeted allocation of risk-management resources.<br>
Identified industries and geographies vulnerable to physical and transition risks, generating actionable insights to support portfolio screening, climate-risk monitoring, and risk mitigation strategies.<br>
Conducted scenario and sensitivity analysis across emissions and risk-exposure assumptions to identify key risk drivers and validate model stability, strengthening the robustness of climate-risk insights for business planning and risk-management decisions.</p>
<p>Analyzed over ₹24,500 crore in transactional data across the two-wheeler loan vertical to derive actionable insights for credit risk assessment and portfolio management.<br>
Developed a Credit Risk Model for a leading NBFC to assess borrower creditworthiness and support risk-based lending decisions.<br>
Built predictive analytics solutions to evaluate and segment borrowers by creditworthiness, enabling differentiated risk assessment and more informed lending strategies.<br>
Identified high-risk geographies and supported portfolio-level risk segmentation to strengthen risk prioritization and exposure management.<br>
Improved credit risk model performance from 78% to 83%, achieving a 5 percentage-point improvement in predictive performance in terms of AUC-ROC and strengthening borrower risk classification.<br>
Evaluated model performance using KS Statistics, AUC-ROC, and Gini Coefficient to assess predictive power, model discrimination, and reliability.<br>
Contributed to reducing the NPA ratio from 5.4% to 4.65%, representing a 0.75 percentage-point reduction (13.9% relative improvement) in the reported ratio through data-driven credit risk assessment and portfolio risk management.</p>
<p>Developed an X-Bucket Collection Optimization Model to identify over 300,000 customers monthly at risk of rolling forward into higher delinquency buckets, enabling proactive collections prioritization.<br>
Designed customer segmentation based on Months on Book (MOB): ≤2, 3–14, 15–24, and ≥25 months to capture differences in repayment behaviour and improve collection targeting.<br>
Applied segment-specific predictive analytics to identify high-risk accounts, enabling targeted follow-ups and more effective allocation of collection resources.<br>
Developed predictive models using Logistic Regression and XGBoost, leveraging Python and SQL to strengthen risk identification and optimise collection strategies.<br>
Evaluated model performance using AUC-ROC, KS Statistics, Gini Index, and Confusion Matrix to assess predictive power and model discrimination.<br>
Optimised collections operations by reducing manual account prioritisation time by an estimated 15–25% and improving collection team productivity by an estimated 20–25% through data-driven customer prioritisation and targeted collection strategies.</p>
<p>Contributed to API development and flight-fare prediction initiatives.<br>
Developed a Random Forest Regression model to analyze flight-fare trends across airlines.<br>
Conducted data analysis to identify priority countries and regions and generate actionable revenue-growth insights.<br>
Worked with Python and SQL for data analysis and predictive modeling.</p>
<p>• Analyzed over ₹24,500 crore in transactional data across the two-wheeler loan vertical to derive actionable insights for credit risk assessment and portfolio management. • Developed a Credit Risk Model for a leading NBFC to assess borrower creditworthiness and support risk-based lending decisions. • Built predictive analytics solutions to evaluate and segment borrowers by creditworthiness, enabling differentiated risk assessment and more informed lending strategies. • Identified high-risk geographies and supported portfolio-level risk segmentation to strengthen risk prioritization and exposure management. • Improved credit risk model performance from 78% to 83%, achieving a 5 percentage-point improvement in predictive performance in terms of AUC-ROC Wand strengthening borrower risk classification. • Evaluated model performance using KS Statistics, AUC-ROC, and Gini Coefficient to assess predictive power, model discrimination, and reliability. • Contributed to reducing the NPA ratio from 5.4% to 4.65%, representing a 0.75 percentage-point reduction (13.9% relative improvement) in the reported ratio through data-driven credit risk assessment and portfolio risk management. </p>
The AI-first Recruitment Operating System that unifies and automates the entire hiring lifecycle for modern enterprises worldwide.
Dr Job FZ LLC. 2026 © All Rights Reserved
