Skills
Certifications
Work Experience
Capital One (contractor through Zenon LLC) | Aug 2023 - present
Lead Data Scientist, Card (Apr 2025 - present) | Senior Data Scientist, Card (Aug 2023 - Mar 2025)
- Boosted marketing campaigns' effectiveness for a $157B credit card portfolio by developing XGBoost models to predict spend and response rates, and strategically targeting high-value prospects, leading to a 7% increase in accounts booked and revenue annually.
- Architected a multi-agentic Gen AI system for the enrichment of stale firmographic and contact data using Google's ADK, enabling precise campaign segmentation, increasing conversion rates by 5%, and reducing marketing effort wastage by 70%.
- Owned the end-to-end RAG experimentation lifecycle - from dataset curation, chunking to reranking, retrieval evaluation - leveraging parallel hyperparameter optimization to achieve 30% accuracy gains while reducing tuning cycles from days to few hours.
- Designed and executed portfolio-wide (10M+ active cardholders) A/B testing experiments to optimize personalized credit limit upgrades and rewards strategies, identifying 15% lift in monthly customer spend.
Discover Financial Services (contractor through Zenon LLC) | Dec 2021 - Jul 2023
Senior Data Scientist, Loans (Nov 2022 - Jul 2023) | Data Scientist, Loans (Dec 2021 - Oct 2022)
- Spearheaded the development of an automated Gen-AI based underwriting software in Python that analyzes application data, agency reports, and triggers GPT-generated rules for optimized process progression, resulting in a 60% reduction in lead time.
- Led cross-functional initiative to upgrade legacy segmentation process for 50M+ prospects, eliminating dependencies and errors, modifying filters to reduce portfolio risk by 25%, and reducing process runtime using PySpark/Dask from 2 weeks to <2 hours.
- Drove transformative improvements in cross-selling through advanced analytics and leveraging predictive models, delivering 30% increase in application volume and 15% improvement in customer lifetime value annually.
- Designed and executed Bayesian experiments on the $128B loan portfolio to identify and filter out high-risk applicants, thus reducing default rates by 10% while increasing approval efficiency.
Carroll School of Management, Boston College | Aug 2019 - Jun 2021
Graduate (PhD) Researcher
- Built clustering-based orthogonalized scores to predict fund manager skill, analyzed its relationship with their performance using multivariate panel data regression, and designed a trading strategy based on it, securing a statistically significant 25% ROI annually.
- Modeled the causal relationship between NLP-based sentiment metrics from 300+ media sources and stock returns to develop predictive buy/sell indicators for portfolio management; achieved a statistically significant portfolio returns of 15% per annum.
- Designed and executed regression discontinuity and instrumental variable regressions, to evaluate the impact of activist investments on firm's market value, and built an investment strategy around it resulting in a 10% increase in ROI annually.
Center for Analytical Finance, Indian School of Business | Jun 2016 - Jul 2019
Senior Research Associate, Data Science (Jul 2017 - Jun 2019) | Research Associate, Data Science (Jun 2016 - Jun 2017)
- Leveraged time-series models and genetic algorithms to study risk patterns in global equity markets and develop investment strategy that optimizes returns and risks, delivering an average annual risk-adjusted return of 20%.
- Processed 7B+ patents/citations from USPTO to create innovation efficiency metrics; analyzed their relationships with equity returns using econometric modeling techniques, built a portfolio based on it generating a 15% ROI annually.
- Built logistic regression models to estimate firm-level credit risks and mitigated exposure to high-risk firms thus minimizing expected annual losses in investment portfolios by 30%.
- Organized statistical inference workshops and provided mentorship to new hires and interns. Excelled in high-pressure environments, managing heavy workloads and meeting tight deadlines, resulting in a 100% rating in annual reviews.