As financial institutions accelerate their adoption of cloud computing and artificial intelligence, some of their most consequential operations continue to depend on fault-tolerant systems designed decades ago. Gudipuri’s published NM4N framework proposes a governed approach to modernising them without compromising resilience.
Behind the polished interfaces of mobile-banking applications and digital trading platforms lies a less visible layer of technology. It processes transactions, moves financial information between institutions and supports operations in which even a brief disruption can carry financial and regulatory consequences.
Some of this infrastructure was designed decades before cloud computing and artificial intelligence became central to enterprise-technology strategies. Yet its age does not necessarily make it obsolete. In many cases, these systems remain in use because they were engineered for an uncompromising requirement: continuous availability.
HPE NonStop is one example of this specialised, fault-tolerant technology. Its architecture has long been used for workloads where an ordinary system interruption may be unacceptable. Modernising applications built for such environments presents a very different challenge from replacing a conventional business application.
“The focus of these systems has never been about bringing in the latest technologies,” says Nithesh Gudipuri, an Associate Director of Technology at Raymond James Financial. “Most cloud environments are focused around scale and the need to accommodate it. With these systems, the focus has always been about availability and uptime.”
Gudipuri’s perspective comes from more than 13 years of experience spanning healthcare, telecommunications and financial services. In his current area of responsibility, he leads 22 technology professionals and holds design, architectural and execution responsibilities for interconnected securities middle- and back-office platforms and enterprise services. The platforms within this operational landscape process an average of approximately 103 million transactions each day.
Earlier in his enterprise-services leadership journey, Gudipuri supported more than 70 services consumed by over 500 business applications. That combination of service architecture, legacy transaction processing and regulated financial operations exposed a recurring modernisation problem: the most important knowledge in an established system is not always contained in its source code alone.
It can also reside in undocumented business rules, operational procedures, data relationships and decisions understood primarily by experienced engineers. A straightforward conversion of code from one language to another may preserve syntax while losing the reasoning that makes the application reliable.
Turning modernisation into a knowledge problem
This challenge is central to Gudipuri’s technical monograph, Neural Modernization for HPE NonStop Systems (NM4N): An AI-Driven Framework for Modernizing Legacy Applications. The work was published through the International Journal of Scientific and Research Publications Monograph and Thesis Series, which carries ISSN 2250-3153.
NM4N approaches legacy transformation as more than a code-generation exercise. It presents an AI-driven framework for examining established HPE NonStop applications, extracting their technical and business meaning, and supporting their transition toward contemporary architectures while retaining human oversight.
The distinction is important. Generative AI can produce software quickly, but speed alone does not establish that the resulting application behaves exactly like the system it is intended to replace. In financial infrastructure, a modernised platform must preserve transaction logic, failure-handling behaviour, data integrity and regulatory controls that may have evolved over many years.
“Even one instance of a system irregularity can result in downstream financial and regulatory implications,” Gudipuri explains. “The way in which you think about building and operating these systems is extremely different from the mindset used for a typical modern distributed environment.”
Rather than treating an older application as a collection of files awaiting conversion, NM4N frames it as a body of operational knowledge that must first be understood. The framework describes a governed process in which artificial intelligence can assist with analysing program structures, identifying dependencies, recovering business meaning and supporting the creation of modern components.
Traceability and expert validation remain essential throughout that process. Proposed outputs must be connected to the source behaviour from which they were derived, allowing engineers to review the reasoning, test equivalence and retain control over consequential architectural decisions.
In this model, AI assists specialists; it does not silently replace their judgement.
Lessons from regulatory transformation
Gudipuri’s work on the framework is grounded in practical exposure to financial-market change. Within his area, he has held design, architecture and execution responsibilities connected with the United States’ transition to T+1 securities settlement, the migration of SWIFT messaging from ISO 15022 to ISO 20022, and preparations associated with regulatory requirements for central clearing of eligible US Treasury transactions.
Each initiative is distinct, but together they illustrate why financial modernisation is rarely isolated to a single application. A regulatory change can require coordinated updates across transaction flows, messaging standards, data models, vendor platforms, controls and operational teams.
The move to T+1, for example, shortened the standard settlement cycle for many US securities transactions from two business days to one. That compression placed greater emphasis on timely processing and coordination across post-trade systems. ISO 20022 presents another kind of transition: moving financial messages toward a richer and more structured data standard while maintaining continuity with established processes and counterparties.
“These programmes require us to transform some of our most critical systems and processes,” Gudipuri says. “In these systems, correctness has to be designed as part of the system. It cannot be treated as an emergent property.”
These regulatory responsibilities are not themselves the subject of NM4N. They provide the professional context behind it: complex financial platforms must continue operating while their underlying technology, interfaces and controls change around them.
Preserving reliability while technology changes
The larger issue raised by NM4N extends beyond any single institution. Organisations across sectors continue to depend on mature systems containing years of accumulated business logic. At the same time, the professionals who understand those environments are retiring or becoming harder to replace.
That creates two connected risks. Institutions may postpone necessary modernisation because the systems are too consequential to disturb, or they may pursue transformation without fully recovering the knowledge embedded inside them.
Gudipuri argues that artificial intelligence can help address this gap when it is applied with governance, traceability and domain-expert review. The objective is not to discard the engineering principles that made fault-tolerant platforms dependable. It is to translate those principles—and the knowledge accumulated around them—into forms that can support the next generation of financial infrastructure.
“The real challenge is making sure that we don’t lose this knowledge faster than we can translate it for the next generation,” he says.
Modern financial markets may increasingly be experienced through cloud services, mobile applications and AI-enabled tools. Beneath that visible innovation, however, are systems built around a much older promise: transactions must continue, records must remain accurate and failures must be contained.
NM4N offers Gudipuri’s answer to a question that will become more pressing as financial technology evolves: how can institutions modernise the infrastructure they depend upon without losing the reliability—and the human knowledge—that made it critical in the first place?
