Security Services Operations (SSO)

Project Overview

The SWIFT network is a global provider of secure financial messaging services, handling thousands of financial messages daily. This project focuses on automating the classification and extraction of key information from these messages using machine learning techniques. By leveraging advanced algorithms, we aim to reduce manual intervention, minimize errors, and improve processing speed.

Key Achievements

• Automated Classification: Successfully developed a system that classifies SWIFT messages into 12 distinct categories, reducing the need for manual sorting.
• Entity Extraction: Implemented machine learning models to extract 23 key entities from messages, such as transaction amounts, dates, and beneficiary details.
• Efficiency Gains: Achieved operational efficiency gains equivalent to 23 Full-Time Equivalents (FTEs), allowing better resource allocation.

Technology Stack

• Programming Language: Python – chosen for its powerful ML libraries and ease of implementation.
• Machine Learning Algorithms: Random Forest – used for high accuracy in message classification.
• Natural Language Processing (NLP): Applied for text analysis and extraction within SWIFT messages.
• DevOps Tools: Azure DevOps – integrated for smooth CI/CD and efficient deployment.