Skip to main content
Customer Cases

FinVolution Group upgrades its development model with Qoder CN: Hundreds of developers generate 33% of new code with AI

By adopting Alibaba Cloud Qoder CN, FinVolution Group significantly improved its coding efficiency and quality, boosting its overall AI-driven development productivity. FinVolution Group is a leading fintech group in China. Its mission is to "Make Finance Better with Technology" using big data and artificial intelligence. The group has a strong Research and Development (R&D) team that focuses on three main areas: credit technology, international business, and tech ecosystem incubation. It has also built strong capabilities in core fields such as credit risk assessment, fraud detection, big data, and artificial intelligence.
image
As a technology-focused company, FinVolution Group excels at creating full-stack fintech products and explores AI technology to enhance its core development capabilities. In recent years, as its business expanded, FinVolution Group's internal software projects grew in complexity and scale, placing higher demands on developers for coding efficiency and quality. The development team faced the following pain points with traditional coding methods:
  • Heavy repetitive work: Writing large amounts of repetitive code, performing manual debugging, and writing code comments consumed valuable developer time and reduced efficiency.
  • Lack of timely optimization suggestions: During daily coding, a lack of immediate code optimization suggestions led to frequent syntax and logic errors. This impacted code quality and required additional resources to correct.
  • High training costs for new employees: New employees spent a significant amount of time reading existing project code and learning coding standards. As a result, the team's well-established coding standards and codebases were not fully utilized.
To address these pain points, FinVolution Group initially developed its own coding assistant, which had some success. However, as business needs grew, the in-house tool showed its limitations, requiring continuous investment in hardware and human resources. Therefore, while continuing to improve its own plugin, the company also began searching for mature, external coding assistant products to further boost development efficiency and quality. When evaluating commercial coding assistant products, FinVolution Group considered international products such as Copilot. However, because of the strict compliance requirements of the finance industry, the group ultimately chose a leading domestic coding assistance tool: Alibaba Cloud Qoder CN. Since its launch in November 2023, Qoder CN has received multiple certifications and honors, including recognition from IDC and the China Academy of Information and Communications Technology (CAICT), and the Best in Show award at the 2024 World Artificial Intelligence Conference. It was also named the only Challenger from China in the 2024 Gartner Magic Quadrant for AI Code Assistants. Currently, the Qoder CN smart coding assistant plugin has been downloaded over 8 million times, generated over 1.5 billion lines of code, and served tens of thousands of enterprises. After learning about FinVolution Group's requirements, the Qoder CN project team held multiple discussions with the FinVolution Group R&D team. They explored in depth the technical implementation of Qoder CN in scenarios such as cross-file contextual code completion, R&D Q&A, and enterprise-level sensitive information filtering. After thorough testing, FinVolution Group evaluated and selected Qoder CN based on the following criteria:
  • Efficient code assistance: Qoder CN uses a model optimized with a specialized code corpus and fine-tuned for mainstream programming languages, such as Java, Python, Go, C/C++, JavaScript, TypeScript, and PHP. It combines this with project-wide awareness of relevant and similar code across files to significantly improve coding accuracy.
  • Support for personalized R&D scenarios: Qoder CN allows enterprises to customize and extend prompt tasks. It also integrates with their private data to enable code completion and enhanced retrieval for R&D Q&A, which meets their personalized development needs.
  • Security and compliance: Qoder CN allows enterprises to customize policies to filter sensitive information, such as passwords and email addresses, from code snippets at the plugin layer. All data is encrypted and protected throughout the process. The official privacy policy guarantees that user code is not stored, which helps build a secure and compliant development environment.
Hundreds of developers at Xinyi Technology are using Qoder CN. With a 70% weekly active user rate, Qoder CN generates 33% of new code and has increased overall development efficiency by 11%, demonstrating a comprehensive improvement in AI-driven development. In the future, Qoder CN will continue to support FinVolution Group by integrating with various tools in the R&D realm, enabling the intelligent upgrade of the company's end-to-end development toolchain.
Overview