A fintech client, rapidly growing fintech company in India, sought a robust and scalable cloud solution to drive real-time insights and better understand customer behavior. The company selected AWS as its cloud partner and leveraged Ancrew’s expertise to establish a secure, durable, and highly scalable data lake infrastructure. To enhance customer engagement and decision-making, client integrated a recommendation engine for personalized service delivery.
About Client
Our client provides a wide range of digital financial services including instant personal loans, credit card management, investment tools, and BNPL (Buy Now Pay Later) options. With operations across urban and semi-urban regions, they partners with banks and NBFCs to deliver fast, secure, and inclusive financial solutions.
Need for Data Lake and Recommendation Engine
Our client handles vast volumes of transactional and behavioral data from its mobile apps, partner portals, and customer service platforms. With millions of users and an ever-growing digital footprint, the company generates and stores extensive data from clickstreams, transactional logs, and relational databases.
The need arose to:
- Enhance business intelligence (BI) capabilities
- Achieve cost-effective and scalable cloud-based storage and analysis
- Execute ETL processes on unstructured and semi-structured data
- Build a recommendation engine to guide users in selecting suitable financial products
Customer’s Journey from Data Pool to Data Lake
To meet its analytics goals, our client used Amazon S3 to build a centralized data lake that captured real-time data from existing sources. AWS Glue was implemented to perform ETL on this data, followed by the use of Amazon SageMaker to train machine learning models that powered personalized recommendations.
Workflow Architecture
The transformation involved a structured workflow:
- Automation of data consolidation from all internal and external sources into a single repository
- Deployment of a VPC on AWS with secure public and private subnets, NAT Gateway, and security configurations
- Site-to-site VPN connectivity with on-premise data centers for secure data transfer
- AWS Kinesis data streams to handle real-time data ingestion
- ETL jobs configured to store both raw and transformed data in Amazon S3
- Machine learning models in SageMaker trained on this data to drive predictive recommendations
- Integration with AWS Redshift for analytics, and QuickSight for interactive BI dashboards
Key Outcomes with AWS QuickSight and Data Lake Implementation
- Reduced Manual Intervention: Automating data ingestion and transformation eliminated the need for manual processes
- Improved Insights: Simplified access to comprehensive insights reduced the complexity and time to value
- Unified Data Repository: Consolidating all datasets into one repository reduced overhead and eliminated the need for multiple database licenses
- Faster, Smarter Predictions: Real-time personalized product suggestions helped users make informed financial choices
- Enhanced Customer Retention: Timely and relevant financial advice improved user satisfaction and reduced customer churn