Designed and tuned risk-based AML thresholds across 10 countries, reducing false positives and improving fraud alert efficiency. Built and validated AML statistical models XGBoost, KNN, clustering) to classify high-risk vs. low-risk customers. Conducted transaction monitoring analytics, fraud pattern analysis, and segmentation using PCA, KMeans, and non-parametric statistical tests. Performed model impact analysis and collaborated with compliance teams to ensure alignment with global AML frameworks. Developed scalable fraud rules engines and optimized alert systems, leading to improved early-stage detection of fraudulent activities. Playing a leading role in one of the HSBC in house AML product where I am responsible for project management, project planning and future development.

Used NLP techniques to extract key biomedical information from research abstracts, improving data tagging efficiency. Created a Python module for dynamic table extraction from PDFs and automated multi-format outputs for client delivery. Conducted statistical analysis of structured and unstructured biomedical data using R and Python. Built an OpenCV-based module to detect medicine ingredients from packaging images. Designed and deployed REST APIs for analytics workflows, scaling performance.

Developed backend statistical algorithms and ML models for software applications. Built a crop recommendation model leveraging soil parameters (pH, Nitrogen, Potassium) and weather features (temperature, humidity). Implemented an object detection system using YOLOv3 OpenCV for realtime applications.