R
Rohit Dhadvai
About
Detail
Gujarat, India
Résumé
Jobs
verified_user
0% verified
AI/ML EngineerStripeOct 2024 - Current (1 year 7 months)• Worked on the development of a Customer Lifetime Value prediction system, analyzing transaction, demographic, and subscription data to provide merchants with actionable insights, improving revenue forecasting accuracy and enabling targeted growth strategies. • Engineered feature sets using Python and SQL, integrating external economic indicators and marketing data to enhance model performance, increasing prediction accuracy by 20% and capturing nuanced customer purchasing behaviors. • Developed the CLVRecNet model combining XGBoost with deep learning embeddings to forecast customer lifetime value and deliver personalized recommendations for upselling and retention, improving merchant engagement and reducing churn risk. • Conduct- CAI/ML EngineerCiti Group IndiaJan 2021 - Jul 2023 (2 years 7 months)• Developed an advanced AI system for real-time fraud detection in credit and debit card transactions and online banking activities, collaborating with fraud operations, data engineering, and compliance teams to gather needs and define KPIs aligned with Citi Group's risk management framework. • Aggregated and preprocessed multi-channel transaction data including cards, net banking, and mobile applications, processing over 99% of incoming transactions using ADF and Databricks while ensuring data accuracy and compliance with Citi's governance ethics. • Engineered fraud-related features capturing transactional anomalies, behavioral patterns, and device fingerprinting, applying methods like time series analysis, clustering, and user profili
Education
verified_user
0% verified
- Master of ScienceGEORGE MASON UNIVERSITYAug 2023 - May 2025 (1 year 10 months)
- Bachelor of Technology (B.Tech)Osmania UniversityAug 2019 - May 2023 (3 years 10 months)
Projects (professional or personal)
verified_user
0% verified
- Customer Lifetime Value Prediction SystemStripeOct 2024 - Current (1 year 7 months)• Designed CLVRecNet, merging XGBoost with embedding layers to predict customer lifetime value and recommend retention strategies, increasing Stripe merchant engagement and reducing churn by delivering actionable insights. • Integrated economic and behavioral data for robust feature engineering, enhancing model performance by 20% and enabling real-time insights via AWS SageMaker dashboard APIs for Stripe's business clients.
- KKnee Osteoarthritis Detection SystemJan 2022 - Mar 2022 (3 months)• Architected a robust system for early detection of knee osteoarthritis in X-ray images using the Xception CNN model with 79% accuracy, integrated into a Django framework, and deployed on Akamai-Linode.
- Real-Time Fraud Detection SystemCitiJan 2001 - Jul 2023 (22 years 7 months)• Developed hybrid AI models combining Autoencoders, Isolation Forests, and XGBoost for fraud detection on imbalanced datasets, improving precision and reducing false positives in high-volume financial transaction streams. • Built Kafka-Spark streaming pipelines with Azure AKS deployment, scoring transactions in real-time and integrating AI alerts into Citi's Security Operations workflows for continuous monitoring and retraining.
- AAnomaly Detection & Behavioral Analytics Pipeline• Applied clustering, time series analysis, and behavioral profiling to extract fraud-related patterns from multi-channel banking data, improving anomaly recognition across digital platforms. • Used Azure Databricks and ADF to automate preprocessing and feature pipelines, ensuring high accuracy and compliance across 99% of Citi's transactional data flow.