The fintech company PRAGMA has announced significant advancements in its AI capabilities, unveiling a new foundation model tailored for financial events, which enhances fraud detection, credit risk prediction, and overall operational efficiency. Historically, the firm relied on siloed machine learning models, leading to slow experimentation and fragmented decision-making; however, with the introduction of PRAGMA, teams can now create new models in days rather than months. This initiative reflects a broader trend among fintech firms towards adopting unified foundation models to minimize duplication of effort and improve performance across functions, supported by the partnership with NVIDIA to optimize deep learning processes.

NVIDIA: NVIDIA develops AI hardware and software platforms used for large-scale model training. The company is collaborating with Revolut to support the deep learning infrastructure behind PRAGMA. This partnership enables efficient scaling of the behavioral foundation model for financial data.
PRAGMA: PRAGMA is Revolut’s proprietary foundation model built specifically for financial services applications. Trained directly on raw financial event streams, it generates reusable behavioral representations that support multiple downstream tasks. In this news, it functions as the company’s unified intelligence backbone to replace fragmented ML pipelines.
Revolut: Revolut is a global fintech company focused on modern banking and financial products. The firm is prioritizing in-house development of advanced AI systems rather than relying on external tools. The news positions Revolut’s work on PRAGMA as a strategic move to accelerate innovation across fraud, credit, and product teams.
Anton Repushko: Anton Repushko authored the post outlining Revolut’s AI strategy and PRAGMA development. He emphasizes the advantages of building custom foundation models over bolting on third-party solutions. His account provides the detailed technical and operational context for the company’s foundation model initiative.

`json
{
“AI Strategy”: “Leading fintech companies are increasingly focusing on unified foundation models, which allow for reduced duplication and better cross-functional performance.”,
“Model Development”: “Organizations are utilizing methods such as freezing pretrained backbones and lightweight adaptation to expedite experimentation and deployment in production environments.”,
“Technology Partnerships”: “Collaboration between fintech companies and providers of AI infrastructure is facilitating the deployment of specialized foundation models in regulated industries.”
}
`

AI tools built by Emerald Force

Built and supported by Emerald Force.

You might also like