Business Objective / Goal
To address the client’s challenges in categorizing income, expenses, payment modes, and merchants across multiple platforms, and to enhance visibility, automation, and insights for smarter financial management.
Solutions & Implementation
- Data Collection: Built an Android app to securely collect user payment SMS data with consent, enabling privacy-compliant financial tracking.
- Message Broker & Processing: Integrated Kafka to handle high-volume SMS traffic and ensure low-latency processing.
- Data Warehouse & Dashboard: Designed a Master Data Warehouse and developed real-time dashboards for visual insights into user income and expenses.
- ML/NLP Engine: Created an ML engine using Python, SpaCy, NLTK, and CRF-suite to classify transactions, map merchants, and categorize financial behaviors.
- Access Control & Admin Panel: Developed ACLs and intuitive dashboards to manage roles, permissions, and user analytics.
- SDK Development: Delivered a plug-and-play SDK replicating the mobile app's core functionality for seamless integration across platforms.
Major Technologies Used
- Python – ML development, data preprocessing, NLP
- Golang – Backend APIs and concurrency support
- Kafka – Message brokering for real-time SMS stream
- RDBMS – Transactional and categorized data storage
- SpaCy, NLTK, Sklearn, CRF-suite, Regex, Pandas, NumPy – AI/ML and NLP pipeline components
Business Outcomes
- Detailed Income & Expense Visibility Categorized financial data with summary dashboards, helping users make better financial decisions.
- Automated Transaction Classification AI/ML pipeline reduced manual tagging, minimized errors, and accelerated insights.
- Enhanced Admin Controls Centralized dashboard for managing users, roles, and access permissions.
- Actionable Insights for Users Forecasts, reminders, and behavior trends drove financial wellness and engagement.
