DenizBank taps Red Hat OpenShift AI to speed up model rollout

DenizBank has incorporated Red Hat OpenShift AI into its workflows to accelerate its AI and machine learning model deployment within its financial services operations.

The Turkish bank, which is the country's fifth largest private bank and employs nearly 15,000 staff, provides a range of banking and financial services including corporate and retail banking, investment brokerage, agricultural banking, SME banking, financial leasing and insurance, as well as technology services through its subsidiaries Intertech and NEOHUB.

Faced with the challenge of slow project development speeds that hindered data scientists working across risk management, marketing and customer relations, DenizBank sought to address operational bottlenecks and create a more efficient environment. Intertech, the bank's IT subsidiary, launched a new AI-supported platform transformation aimed at supporting around 120 data scientists by improving the reliability and performance of their workflows.

DenizBank selected Red Hat OpenShift AI, which allows for the management of AI and machine learning lifecycles across both on-premise and public cloud environments. The bank had previously adopted Red Hat OpenShift in 2023 for its containerisation needs. Integration of Red Hat OpenShift AI enabled automation and standardisation of data science pipelines, the development of self-service capabilities, and the scaling of model serving to meet organisational needs.

By using Red Hat OpenShift AI, DenizBank reports that its data scientists are now able to build tailored AI model development environments with plug-and-play connectivity for data sources, which significantly streamlines the process of selecting and validating data for training, fine-tuning and validation. As a result, the time required for new microservices models to go from development to market has been reduced from several days to just 10 minutes, according to the bank.

Integration with new hardware accelerator dashboards allows for optimised use of GPUs, with Red Hat OpenShift AI scaling GPU slices for machine learning models as needed. This improves resource utilisation, increases flexibility, and enables more workloads to run simultaneously without the need for additional hardware.

Okan Çetinkaya, Chief Data Officer and Chief Administrative Officer at DenizBank, said, "Today, our data scientists play a key role in supporting decision-making across our bank with data, one of our most valuable assets. With Red Hat OpenShift AI, they can build and deploy robust models in a streamlined, standardized and flexible development environment, while integrating with a wide range of open source tools. We are pleased to empower DenizBank's talented team of data scientists through our collaboration with Red Hat, supporting their drive for innovation and excellence."

The bank highlights that the deployment of Red Hat OpenShift AI has improved the quality and speed of model development and operationalised both generative and predictive AI models with greater ease. The self-service model has empowered data scientists to focus on high-value tasks, further standardising quality control and expediting deployments across departments.

Haluk Tekin, Country Manager, Türkiye and CIS at Red Hat, stated, "We are proud to be able to help DenizBank accelerate time-to-market with the power of AI, as they pioneer innovation in Turkey's fast-paced financial services sector. Red Hat OpenShift AI pairs the power of open source innovation with a hardened, enterprise-ready AI platform, enhancing the ability of DenizBank to build employee trust and empowerment around AI applications today and for the future."

Ömer Uyar, Chief Executive Officer at Intertech, commented, "Adding Red Hat OpenShift AI to our arsenal has significantly enhanced our data scientists' autonomy and improved our operational standards. This perfectly aligns with our commitment to accelerating innovation while maintaining a high-quality framework. We highly value the impact of Red Hat OpenShift AI's advanced capabilities and automation in empowering our data scientists to achieve greater efficiency and innovation. At Intertech, we aim to realize much more measurable improvements in both the speed and quality of AI-driven projects in the near future."

DenizBank also reported enhancements in prediction models used to determine customer credit loan eligibility and improvements in fraud prevention systems, with the AI-powered systems now able to identify potentially problematic financial patterns more quickly.

With the establishment of company-specific standards and automated pipelines, DenizBank's future plans involve the migration of all data scientists and models to the Red Hat OpenShift AI platform. The organisation is also focusing on further optimising compute resource utilisation, including CPUs, GPUs, and memory, as well as enhancing model inferencing speed and capability using Red Hat AI solutions.