Explore My Projects

PaymentGard-AI

Built an ML model predicting SaaS payment defaults with 87% accuracy (AUC: 0.89). Identified $142K/month at-risk revenue, reducing payment-related churn by 38%. Implemented an early-warning system, helping businesses recover $1.7M annually through proactive interventions.

Website-Performance-Impact

Analyzed 2.5M+ user events to reduce 63% cart abandonment ($12.8M/month loss) through ML-driven recovery. Built an XGBoost model (AUC: 0.87) predicting high-risk abandoners, enabling 92% targeted retention efforts. Designed AI-powered remarketing and discounting flows (+35% conversion lift).

Customer-Analytics-and-Business-Impact

Sentiment-Driven-Product-Launch-Strategy-Solution

Analyzed 500K+ Amazon reviews using NLP (TextBlob, LDA) to optimize product launch strategy. Built a sentiment-driven predictive model (87% accuracy) to flag high-risk product launches. Reduced negative reviews by 32%, improving satisfaction by 18%.

Developed ML models (85% accuracy) to identify high-value customers and reduce churn by 15%. Optimized CLV by improving targeted retention campaigns. Identified top-performing products driving 40% of revenue, refining marketing strategies.

Predicting-Viral-Trends-on-Social-Media

Built an ML model (85% accuracy) analyzing YouTube trends for early viral content detection. Identified top-performing content categories (tech, fashion) with 10-12% engagement rates. Used Prophet time-series forecasting for virality predictions 30-90 days ahead.

Credit-Card-Chargebacks-Prediction

Developed a fraud detection model (AUC-ROC: 98%) reducing fraudulent chargebacks by 30%, saving $5M annually. Achieved 92% precision and 85% recall on 284K+ transactions. Automated real-time fraud monitoring, reducing manual reviews by 40%.

Payment-Delays-Cash-Flow-Optimization

Analyzed 1,000+ invoices, identifying 10 delinquent accounts causing 30% of payment delays. Reduced late payments by 32% (22 → 15 days) through targeted interventions. Implemented early-payment discount programs.

Identifying Ghost Customers in Subscription Services

Built a Random Forest model (85% accuracy) to detect ghost customers and reduce churn by 15%. Identified key churn predictors (low login frequency, engagement <20). Designed win-back campaigns (+20% renewal rate) and personalized email flows (30% open rate).

E-Commerce Cart Abandonment Analysis

Analyzed 2.5M+ user events to reduce 63% cart abandonment ($12.8M/month loss) through ML-driven recovery. Built an XGBoost model (AUC: 0.87) predicting high-risk abandoners, enabling 92% targeted retention efforts. Designed AI-powered remarketing and discounting flows (+35% conversion lift).

Decoding-Black-Friday

Analyzed 550K+ transactions to optimize Black Friday discounts (22-38% range), increasing conversions by 37.5%. Built an XGBoost model (89.2% accuracy) to identify key customer segments for dynamic pricing. Implemented real-time price adjustments across 15K+ SKUs.