OceanoBe Technology

OceanoBe Technology OceanoBe | Time to Be Unstoppable |
Building software solutions for a better world

From 14 November, Swift will reject CBPR+ payment messages that carry fully unstructured postal addresses, and in June a...
12/11/2026

From 14 November, Swift will reject CBPR+ payment messages that carry fully unstructured postal addresses, and in June around 60% of debtor addresses in pacs.008 traffic were still free text. Most banks now need a service that structures those addresses reliably, both for historical customer data and for live payments.

Here's our walk through building that service in Java: a rules layer that resolves the predictable majority from postcodes, IBAN country prefixes and per-source grammars; an LLM, via Spring AI, that extends coverage to the messy tail; and a deterministic verifier that rejects any town name not literally present in the source text, because a model given a postcode and a street will happily infer a city that was never written down. Anything uncertain goes to a human repair queue rather than to the network.

https://oceanobe.com/news/the-14-november-iso-20022-deadline/2045

Anthropic's announcement that Claude Code can analyse and translate COBOL moved IBM's stock down more than 13% in a day....
10/11/2026

Anthropic's announcement that Claude Code can analyse and translate COBOL moved IBM's stock down more than 13% in a day. That's the market betting AI-assisted legacy modernisation is real. It's not the same as knowing whether the Java it generates is safe to run on your core ledger.

https://oceanobe.com/news/claude-and-copilot-on-legacy-cores/1993

Ship a React Native banking app in one country and accessibility, localisation are two checklist items. Ship the same co...
09/11/2026

Ship a React Native banking app in one country and accessibility, localisation are two checklist items. Ship the same codebase into eight countries and they become a combinatorics problem: every screen, times every locale, times every screen reader, times every release. This one's for the React Native developer who owns that matrix and has stopped pretending a human can hold all of it in their head.
https://oceanobe.com/news/accessibility-and-localisation-at-scale-for-multi-country-banking-apps/1992

Ask most frontend teams where LLM guardrails live and you'll get the same answer: the backend. Prompt engineering, retri...
06/11/2026

Ask most frontend teams where LLM guardrails live and you'll get the same answer: the backend. Prompt engineering, retrieval pipelines, output classifiers, tool-call permissions. These are all backend concerns, all somebody else's problem. That's true right up until the moment a hallucinated fee amount renders in a chat bubble with the same visual authority as a verified balance, or a cleverly worded paste triggers an action nobody asked for. At that point it's not a backend incident. It's whatever the customer saw on their screen, and the screen was built by the frontend team. ➡️ https://oceanobe.com/news/guardrails-in-the-ui-layer/1991

Requirements haven't changed. What's changed is who, or what, reads them first. When a developer picked up a ticket, the...
03/11/2026

Requirements haven't changed. What's changed is who, or what, reads them first. When a developer picked up a ticket, the spec was a starting point for a conversation. When an agent picks up the same ticket, the spec is the conversation, and whatever it doesn't say gets decided by a model with no stake in getting it right and no instinct to ask. For a BA, that's not a tooling change. It's a change to the actual skill.

https://oceanobe.com/news/writing-specs-an-agent-can-execute/1990

Most DORA compliance content is written for risk and legal teams, and reads that way: obligations, deadlines, templates....
02/11/2026

Most DORA compliance content is written for risk and legal teams, and reads that way: obligations, deadlines, templates. That's fine for the register of information. It's not much use when you're the engineer who has to make a live production system survive a red team, or make a vendor integration exit-able on paper actually be exit-able in code.
This one's for that engineer: https://oceanobe.com/news/designing-for-dora/1989

For every $1,000 that moves from a bank account into a third-party stablecoin, only about 15% comes back to the banking ...
29/10/2026

For every $1,000 that moves from a bank account into a third-party stablecoin, only about 15% comes back to the banking system as reserves.

That single stat is why stablecoins stopped being a compliance footnote and became a balance sheet problem, and, with the GENIUS Act's January 2027 deadline approaching, an infrastructure problem with a clock on it.

Here's our layout of technical questions treasury teams should be putting to IT before committing to a stablecoin or tokenized deposit rollout, on reserve custody, settlement reconciliation, intraday liquidity, compliance monitoring, and vendor resilience.

Read it here: https://oceanobe.com/news/stablecoins-and-bank-balance-sheets/1988

A partner's traffic spike, bug, or compromised credential shouldn't become your incident. That's the design principle be...
28/10/2026

A partner's traffic spike, bug, or compromised credential shouldn't become your incident.
That's the design principle behind API gateways for embedded finance, where non-bank partners plug directly into your core banking capabilities: account creation, payments, KYC, balance inquiries. The gateway isn't just routing traffic. It's the enforcement point for every trust decision in the partnership agreement.

OceanoBe designs banking integration architecture with partner risk isolated by default, not managed after an incident. Here is our perspective on getting this right with Spring Cloud Gateway (with code snippets): https://oceanobe.com/news/api-gateway-design-for-embedded-finance/1984

Credit scoring is a high-risk category by default under the EU AI Act's Annex III. This means explainability, audit trai...
26/10/2026

Credit scoring is a high-risk category by default under the EU AI Act's Annex III. This means explainability, audit trails, and human oversight can't be bolted on after a model ships. Our article breaks down how to build interpretable credit scoring systems from the model selection stage onward: from documentation-as-code to designing human oversight hooks that enable real intervention, not rubber-stamping.

https://oceanobe.com/news/building-credit-scoring-models-that-satisfy-the-eu-ai-act/1983

A "centralized" data lake for a banking group spanning multiple jurisdictions often ends up re-fragmenting anyway, just ...
21/10/2026

A "centralized" data lake for a banking group spanning multiple jurisdictions often ends up re-fragmenting anyway, just without the ownership clarity a mesh architecture would have given it from the start. The real decision isn't data mesh vs. data lake as an architectural preference, it's an honest inventory of where your subsidiaries' governance requirements genuinely diverge, and where they don't. We break down the governance and latency trade-offs for multi-entity banking groups from the perspective of a Java engineer who's built both models.

https://oceanobe.com/news/data-mesh-vs.-centralized-data-lake/1981

Address

Iasi

Alerts

Be the first to know and let us send you an email when OceanoBe Technology posts news and promotions. Your email address will not be used for any other purpose, and you can unsubscribe at any time.

Contact The Business

Send a message to OceanoBe Technology:

Shortcuts

Share