Enabling financial inclusion with faster loans
Reduction in cloud infrastructure costs
Faster data processing shortens loan underwriting time
Financial institutions in Korea are looking for ways to streamline the loan process, provide more personalized loan products, and utilize advanced technologies like big data and AI to make more informed lending decisions. Finda is a data-driven lending platform offering over 200 loan products from 66 banks and financial companies. To enhance data governance and address issues with multiple analysis environments, Finda turned to Databricks. The company uses Databricks solutions to unify distributed data analysis environments and tools to attain data governance for personal information management and better visibility of integrated data. Finda can also operate its infrastructure efficiently by taking prompt action on operational issues.
Inefficient operation of a legacy infrastructure
The lending industry has undergone a revolutionary transformation in Korea, thanks to rapid advancements in analytics and automation. Through improved risk management, faster underwriting times and easier credit access, this transformation has significantly impacted the Korean economy. With a mission to reduce barriers to loans and democratize credit access, Finda is leading the charge in driving greater financial inclusion.
One of the fastest-growing startups in the fintech industry, Finda leverages loan-related data captured through a MyData license from 324 financial companies to provide targeted lending services to customers. However, it struggled to support spikes in data volumes and an increase in data users. Additionally, its complex data environment was made up of different analysis systems used for various analysis demands, making it more difficult to extract data insights and value for its customers. The company also experienced frequent application outages due to scalability issues that limited its ability to respond to sudden increases in users or operational activity. It became difficult to manage data engineering activities such as table creation, modification, and deletion in the service database, which was used for back-end services — absorbing valuable resources and impacting SLAs.
Core to this issue was Finda’s legacy data warehouse. The legacy system was initially established to minimize compliance risk and simplify ease of access to its infrastructure. However, the system was inefficient in managing storage, resulting in runaway operating costs. Additionally, the system required constant maintenance to synchronize the data catalog on both storage environments. “To address issues with manageability, operational inefficiency and cost control, we recognized the need to modernize our data infrastructure in order to unify data, analytics and AI,” said Hui Seo, Chief Technology Officer, Finda.
Increasing operational efficiency and minimizing risk with lakehouse architecture
Finda implemented the Databricks Data Intelligence Platform, a solution that combines the best elements of data lakes and data warehouses to deliver the reliability, strong governance and performance of data warehouses with the openness, flexibility and machine learning support of data lakes. With a modern lakehouse architecture, Finda can now manage all aspects of its data, analytics and AI efforts in a single, unified view. Infrastructure management has been simplified as silos between disjointed analytics services have been eliminated, and integration with an internal GitHub environment makes it easier to share analysis results with team members, boosting collaboration and cross-team productivity.
With Delta Lake and Spark Streaming, Finda has greatly improved data pipeline performance at scale while reducing operational costs by no longer needing to operate its legacy data warehouse and eliminating the need to duplicate data. Databricks Unity Catalog allows Finda to establish data governance through fine-grained access controls for all its users. With these capabilities, Finda has established a data analytics environment that restricts access to customers’ personal information and minimizes compliance risks.
“Using Databricks Data Intelligence Platform, we have created a unified data environment that ensures data governance for personal information management and visibility of integrated data. Databricks has also streamlined our infrastructure, maximizing operational efficiency and reducing overall costs,” said Hui Seo, Chief Technology Officer, Finda.
Improving customer experiences with faster loans
With Databricks Data Intelligence Platform, Finda can now process loans faster and help customers make more-informed business decisions. With a modern lakehouse architecture in place, Finda has also experienced significant cost savings — reducing monthly licensing fees and cloud resource costs by an estimated 40%. From a performance standpoint, it has reduced the time between record creation and storage from eight minutes to two minutes. As a result, the company can conduct analysis more efficiently and deliver high-value data products to customers more quickly.
Looking ahead, Finda now has the data foundation and confidence to continue its efforts to provide accurate financial diagnoses to its customers based on better data while building innovative data products that can better serve clients.
SeoHee concludes, “Databricks is foundational to our mission to modernize lending with data, analytics and AI. With unified data analysis and streamlined infrastructure, we can now take prompt action on operational problems, stay cost-efficient and deliver faster loans to our customers.”
