Automated the data preparation and massaging process by developing an orchestration framework, resulting in significantly reduced processing time and improved efficiency by 5 times than earlier process.•Analysed business requirements and implemented efficient data pipelines by ingesting data into SQL, enabling seamless integration with downstream processes.•Applied various aggregations and feature engineering techniques prior to model building, enhancing model accuracy and predictive power.•Developed robust machine learning ensemble architectures and neural networks model on Keras and Pytorch to detect fraudulent cases, achieving high performance and accuracy.•Utilised confusion matrix, ROC-AUC curve, and scikit-learn's evaluation metrics as an evaluation metric, providing a clear understanding of model performance and guiding future model development.•Utilised the shapely tool to conduct feature analysis and identify key features with high sensitivity towards the model, leading to a deeper understanding of the data and model performance.•Automated every process as pipelines for data preprocessing, modelling, and monitoring, by using MLOPS architecture in Amazon AWS Sagemaker and Amazon S3 efficiently.