EXL helps global financial institution optimize Big Query codes for GCP migration
The client, a globally recognized financial institution, is a leader in the financial services industry handling immense volumes of data daily. Operating within a highly competitive sector, they embarked on an enterprise initiative to migrate from their legacy systems to a modern, cloud-based infrastructure (GCP). However, the technical complexities of code conversion and optimization posed a significant challenge that could impact not only system performance but also operational efficiency. Recognizing the need for a partner well-versed in advanced AI-guided solutions, the client engaged EXL to streamline this transformation.
Challenge
The client set out to migrate over 10,000 HQL-based Big Query codes to the GCP cloud platform. However, their existing conversion tool fell short in several critical areas:
- Optimization challenges: The converted codes were suboptimal, failing to maintain logical coherence and underperforming in terms of compute efficiency.
- High computational costs: Without proper optimization, the codes required longer compute times and consumed excessive memory, leading to inefficiencies.
- Manual intervention requirements: Post-conversion, developers often needed to manually refine and optimize code, consuming valuable resources and increasing the risk of errors.
The scale and complexity of this migration demanded a more robust, intelligent solution that could enhance performance while reducing manual effort.
Code Harbor™ was designed to process large code repositories efficiently in a single pass while maintaining contextual integrity, making it ideal for enterprise-level transformations
Solution
EXL introduced its proprietary Code Harbor™ platform, a generative AI-powered multiagent solution, to address the challenges faced by the client. The solution leveraged state-of-the-art AI models combined with expert customization for precise code conversion and optimization. Key aspects of the solution included:
- Multi-agent framework: Code Harbor™ employed specialized AI agents for different stages of the migration process, ensuring optimization across syntax, logic, language nuances, and code readability.
- Synthetic data for testing: To ensure accuracy, EXL generated synthetic data to rigorously test code outputs and validate results before implementation.
- Automated optimization: The solution transformed HQL-based codes into highly efficient, GCP-compatible Big Query formats, eliminating the need for post-conversion manual changes.
- Scalable design: Code Harbor™ was designed to process large code repositories efficiently in a single pass while maintaining contextual integrity, making it ideal for enterprise-level transformations.
EXL’s team worked closely with the client to deploy Code Harbor™ seamlessly, ensuring minimal disruption to business operations while meeting all compliance and security requirements.
Results
The deployment of Code Harbor™ delivered significant benefits for the client:
20-30% reduction in compute time and memory usage
The AI-driven optimizations enhanced computational efficiency, leading to reduced operational costs.
Streamlined query plans
A marked reduction in code query plans improved overall system performance.
No manual adjustments require
All optimized codes were deployment-ready, saving the client significant time and resources.
Improved governance and documentation
Detailed documentation was created for all optimization changes, providing clarity and supporting future enhancements.
These outcomes not only addressed the immediate challenges of the migration project but also set the foundation for long-term operational scalability and efficiency.