Subhin Antony Email & Phone Number

Bengaluru, Karnataka, India

Bengaluru, Karnataka, India

• Integrated data from various sources and developed around 1,000 features across four critical areas: application data, repayment patterns, delinquency trends, and credit bureau information.• Developed an Early Warning System Model using Random Forest and Boosting machine learning algorithms.• Successfully created a model that detected 80% of payment bounces by focusing on just the top 20% of high-risk debtors.• Developed over 1,000 features from consumer and commercial bureau data, financial statements, and banking information.• Selected key variables based on Information Value (IV) & business insights & applied Weight of Evidence (WoE) transformation for binning.• Identified critical features, such as a Z-score, to detect fraudulent patterns in financial transactions.• Applied techniques like cost-sensitive learning, resampling, and ensemble methods to address the challenge of imbalanced datasets in fraud analytics.• Combined diverse data sources, including Google search trends, marketing KPIs (intent, awareness), and flight capacity data, to develop a comprehensive dataset.• Developed an OLS Regression Model incorporating seasonality components to forecast monthly hotel visits in Dubai, delivering actionable insights for strategic planning.• Developed an AI-powered solution using OpenAI's Whisper for multilingual transcription and GPT-4 for translation and context analysis, 30% reduction in processing time for multilingual customer interactions.• Used Paddle and Tesseract OCR with OpenCV to process 500+ documents weekly, achieving 95% accuracy in extracting text and speeding up document handling.

Bengaluru, Karnataka, India

• Identified fraudulent patterns in financial transactions through data-driven insights.• Tackled imbalanced classification problems in fraud analytics using cost-sensitive learning, resampling, ensemble learning, and auto-encoders.• Deployed an interactive fraud analysis dashboard using Plotly, improving fraud detection and user experience.

Pune, Maharashtra, India

I worked on a diagnostic system that allows maintenance managers to anticipate the risk of future failures and identify asset failure factors that could have an influence on business operations.

Pg Diploma, Data Science

Bachelor Of Engineering - Be, Electrical, Electronics And Communications Engineering