Intro

Jeff App is a data-driven financial supermarket for emerging markets. Their platform connects millions of underbanked borrowers with hundreds of lenders, using alternative data to assess creditworthiness and unlock financial access. Operating in multiple countries across Asia and Latin America, Jeff App has processed over 100 million applications, demonstrating massive traction and product-market fit.

As the company scaled, its engineering team faced significant platform delivery bottlenecks. The infrastructure and processes that worked for a small startup were now hindering their ability to innovate, ship features and maintain a stable service for their rapidly growing user base. They needed a partner to help them build a modern, scalable and efficient cloud platform. They chose GOStack.

The Challenge

Jeff App’s rapid growth exposed the limitations of their early-stage infrastructure. The engineering team was struggling with several critical challenges that slowed down development and increased operational risk.

Delivery Bottlenecks: The existing CI/CD process was slow, unreliable and required significant manual intervention. This created a major bottleneck, delaying feature releases and making it difficult to respond quickly to production issues.

Lack of Developer Autonomy: Developers had limited control over their environments. They depended on a central operations team for infrastructure provisioning and deployments, which stifled productivity and created a culture of dependency.

System Instability: The platform suffered from stability issues under increasing load. Without proper observability and automated scaling, the team was constantly fighting fires and reacting to incidents rather than proactively improving the system.

Inefficient Cost Management: Cloud costs were rising without clear governance or visibility. The team lacked the FinOps practices needed to optimise spend, rightsize resources and ensure that infrastructure costs scaled efficiently with business growth.

Our Solution

GOStack implemented a full platform modernisation programme focused on establishing a self-service development platform, a GitOps operating model and a robust, scalable AWS architecture.

Infrastructure as Code (IaC): We established a strict IaC discipline from the outset. All AWS infrastructure was defined and managed using Terraform. This included the Amazon EKS clusters for container orchestration, Aurora PostgreSQL and DynamoDB for the data layer, Amazon ElastiCache for caching, and the serverless components using AWS Lambda and Amazon API Gateway. This ensured every environment was reproducible, auditable and consistent.

GitOps with ArgoCD: The entire deployment workflow was rebuilt around GitOps. We replaced the legacy deployment process with a modern CI/CD pipeline using GitHub Actions for continuous integration and ArgoCD for continuous delivery. Every change to an application or infrastructure component now goes through a Git-based pull request workflow, with ArgoCD automatically reconciling the state of the Kubernetes cluster with the configuration defined in Git. This eliminated manual deployments and provided a full audit trail for every change.

Self-Service Development Platform: The new architecture empowered Jeff App’s developers with true self-service capabilities. They can now spin up preview environments for feature branches, deploy services to production and access logs and metrics independently. This autonomy has dramatically improved developer productivity and ownership.

Modern Data & ML Platform: We built a cohesive data and machine learning platform to support Jeff App’s core business. Raw data from multiple sources is ingested into Amazon S3. AWS Glue is used for ETL jobs to process and transform this data. Amazon Redshift serves as the central data warehouse for analytics, while Amazon Athena provides ad-hoc querying capabilities directly on S3. For their data science and credit scoring models, AWS SageMaker was integrated to provide a fully managed machine learning environment, enabling the team to build, train and deploy models at scale.

FinOps and Cost Optimization: We introduced cloud financial management best practices, including comprehensive resource tagging, cost allocation reporting and automated scaling policies. This gave the team the visibility and tools needed to manage their cloud spend effectively, leading to a 25% cost reduction.

Results and Benefits

The partnership with GOStack transformed Jeff App’s engineering capabilities and provided a stable foundation for future growth.

• Improved Developer Productivity: With a self-service platform and automated CI/CD pipelines, developers can now ship features faster and with greater confidence.

• Cost Optimization: The implementation of FinOps practices and architectural improvements resulted in a 25% reduction in monthly infrastructure costs.

• Enhanced System Stability: The new platform is more resilient, scalable and observable, significantly reducing the frequency and impact of production incidents.

• Enabled Self-Service: The engineering team is now fully empowered to manage the entire lifecycle of their services, fostering a culture of ownership and innovation.