Tanika Gupta

Designed and implemented an agentic GenAI underwriting system (Loan Lens AI) that won the 2025 Financial AI Hackathon, personally recognized by Andrew Ng for innovation and technical excellence, Led a cross-functional team to win the Barry Callebaut GenAI Hackathon, surpassing 11 competing vendors with an innovative AI solution that generated $1.3 million in business value., Led a team of AI engineers and data scientists in developing and deploying a Gen AI-powered product recommendation system, automating product matching based on customer briefs. This streamlined operations, reducing manual effort and enhancing sales efficiency by 25%, leading to a 15% increase in conversion rates., Provided technical leadership in designing and launching a Gen AI-driven sales intelligence chatbot, enabling natural language queries on sales data. This accelerated decision-making, reduced query response time by 60%., Led the development of an AI-powered promotion optimization system, analyzing historical promotions and external factors to quantify uplift and recommend data-driven promotion strategies, leading to improved marketing effectiveness. This resulted in a 20% increase in marketing ROI., Optimized product pricing strategy for a leading CPG giant by developing a ML model to analyze price, distribution, and macro-economic factors leading to a 4% increase in retail sales and a 2% improvement in gross margins., Developed a scalable demand forecasting solution for a top beverage company, cutting model runtime from 35 hours to 2 hours using advanced time series techniques with macroeconomic and retail data.

Conducted pre-campaign data analysis for targeted marketing initiatives, leveraging statistical modeling and data mining techniques to optimize credit card upgrade and home loan campaigns. Analyzed customer behavior patterns to enhance campaign effectiveness and conversion rates.

Single-handedly driven the end-to-end modelling initiative for the first-generation POS Fraud Models for Amex Prepaid Cards Bluebird and Serve using BIG Data modelling technique GBM (Gradient boosting machine). Simulated complex CAS variables and improved the dependent variable definition for the model., Developed an innovative PageRank-based merchant ranking system to identify influential merchants within a region, enabling targeted bundled offers. This solution optimized merchant partnerships and improved customer engagement., Designed and implemented high-performance fraud detection rules, leading to a $300 K annual reduction in debit card fraud for the Bluebird platform., Successfully completed the American Express Modeling Training Curriculum in Q1 2014, deepening expertise in fraud detection modeling and advanced predictive analytics.

Spearheaded the development of a cutting-edge fraud detection model for Chase’s digital products, leading a team of AVPs and associates. The model resulted in a 30% reduction in fraud incidents., Architected and deployed a GPU-accelerated XGBoost model in Python on a Hadoop cluster, significantly reducing debit card fraud and delivering $20M in annual savings., Pioneered an automated fraud claims tagging system, leveraging advanced NLP techniques (BOW, TF-IDF, Topic Modeling) to cut manual processing time by 50%, and increase operational efficiency, driving cost savings and improved compliance.

Filed three patent applications for MasterCard, driving innovation in AI-driven financial solutions and strengthening the company’s intellectual property portfolio., Developed and deployed advanced image recognition models to analyze facial emotions, enabling banks and merchants to enhance customer interactions., Designed a predictive attrition model for CITI India, leveraging advanced machine learning techniques to forecast customer churn within a 3-month window. The model successfully reduced attrition rates from 14% to 10%, enhancing customer retention strategies and revenue growth.