Financial AtlasTo the mapУкраїнська

SERVICE INFRASTRUCTURE

Technology & Infrastructure

The systems where money actually lives: no account or payment works without them.

Open on the interactive map →

This page is the text version of one zone of the interactive map.

Its role in the system

The digital infrastructure on which money runs: the overwhelming majority of it is records in the systems of banks, payment networks and depositories. Technology companies and IT teams keep those systems switched on, defend them against attacks and change the way financial services are built and delivered — from mobile banking to blockchain.

Functions

  1. Core banking and processing — the systems that keep accounts, accrue interest and process payments (Temenos, FIS, Oracle FLEXCUBE, the cloud-native Thought Machine, or in-house builds); at insurers and funds — policy administration and asset accounting systems.
  2. Cybersecurity and anti-fraud — protection against attacks, standards for card data (PCI DSS), detection of fraudulent transactions in real time.
  3. Cloud and infrastructure — moving systems into the cloud (AWS, Azure, Google Cloud), fault tolerance, backup data centres.
  4. Data, AI and ML — credit scoring, fraud detection, personalisation, back-office automation, generative AI in customer service; since the end of 2025 the payment networks have been opening their rails to AI agents that complete a purchase on a person's behalf (Visa Intelligent Commerce, Mastercard Agent Pay).
  5. RegTech and SupTech — automation of compliance (KYC, transaction monitoring, reporting) inside institutions, and analytical tools for the regulators themselves.

What makes it special

  • How it earns: software licences and subscriptions (SaaS), transaction fees, implementation and support. For a financial institution IT is one of the largest cost lines and at the same time a source of advantage over competitors.
  • The legacy problem: many large banks still run on systems written decades ago (mainframes, COBOL). Replacing them is like changing an aircraft engine in flight, so migrations take years and often end in outages; some banks are instead moving the core into the cloud step by step. 🇯🇵 Japan is the extreme case: bank cores still live on mainframes, and dozens of regional banks share one contractor's platform — NTT DATA has run its joint centre for regional banks on its own BeSTA core since 2004 and only plans to move it into its own cloud by 2028. Mizuho showed the price of a mistake: after switching to the new MINORI system the bank went through eight failures during 2021 — ATMs swallowed cards, branches stopped serving customers — and the FSA issued business improvement orders.
  • The supplier under supervision, and concentration risk: technology companies usually hold no financial licences, but regulators require institutions to control their suppliers. In the 🇪🇺, DORA has applied since 17 January 2025 — digital resilience rules for the financial sector that also introduced direct oversight of critical IT providers, cloud providers among them. The reason is simple: if thousands of institutions run on one cloud or with one vendor, an outage or a breach there hits the whole system at once.
  • 🇨🇳 China: finance inside a super-app: everyday payments run through two super-apps — Ant's Alipay and WeChat Pay — and the platforms themselves rest on home-grown technology: the distributed database OceanBase, built inside Ant and spun off into a separate company in 2024. It is precisely that weight that made the authorities force big tech's financial businesses to operate under banking rules: in July 2023 Ant was fined around seven billion yuan, and with that the regulators declared the sector's "rectification" complete. The model is instructive: technology here does not serve the bank from outside — it becomes the place where people keep their money and pay.
  • Blockchain and distributed ledgers (DLT): the idea is simple — instead of every party keeping its own record and then reconciling it with the others, everyone works from one shared ledger, and a smart contract executes the terms automatically: transferring the asset and paying for it become a single action, and reconciliation disappears. Tokenisation is the same bond, fund unit or bank deposit held as an entry in such a ledger. What genuinely works: 🇨🇭 Switzerland has given banks a wholesale digital franc for settling digital bonds on the SDX exchange since the end of 2023 (the pilot runs to at least mid-2027), J.P. Morgan runs tokenised deposits in Kinexys, and the 🇺🇸 GENIUS Act (July 2025) and 🇪🇺 MiCA put stablecoins under rules with full reserving. What does not: the 🇪🇺 DLT Pilot Regime has applied since March 2023, yet by mid-2025 only three infrastructures had been authorised — ESMA blames the lack of access to central bank money, weak interoperability with ordinary systems and caps that are too tight.
  • Open banking: with the customer's consent, a bank must open access to their account to other licensed services through an API — to show the balance and statement (account information) or to initiate a payment without a card. The customer gives and withdraws the consent, and the third-party service holds its own licence and supervisor. 🇪🇺 the rules were set by PSD2 (since 2018) together with strong customer authentication; next come PSD3 and the PSR regulation, agreed in 2026 and applying in 2027–2028, while the FiDA regulation on access to wider financial data is still being negotiated. 🇬🇧 the model was imposed: after an order from the competition authority the largest banks opened their APIs, and in 2025 that produced 351 million payments and 16.5 million user connections (not unique people). 🇺🇸 the rule on personal financial data rights was adopted in October 2024, but a court blocked its enforcement and the agency is rewriting it — a rule on paper rather than in practice.

Why it matters

When technology fails, money "disappears" from the screens: clients cannot pay, banks cannot process transactions, and a coding error or a cyberattack inflicts enormous losses within minutes and undermines trust in the institution.

🇬🇧 TSB, April 2018

The British bank was moving client data from an old IT platform to a new one; after the migration online and mobile banking, branches and phone channels failed on a massive scale, and full recovery dragged on until December. Regulators fined the bank 48.65 million pounds for poor management of the programme and of the IT supplier's risks.

🇺🇸 Knight Capital, August 2012

During an update of trading software, forgotten old code sent more than 4 million orders to the exchange in 45 minutes; the loss exceeded 460 million dollars and one of the largest US brokerage firms was left on the brink of collapse. The system had sent automated warnings before trading even opened — nobody acted on them.

🇺🇸 CrowdStrike, July 2024

On 19 July 2024 a faulty content update from the security vendor CrowdStrike crashed Windows computers around the world within minutes — on Microsoft's estimate about 8.5 million machines. Banks, payment processors, exchanges, airlines and hospitals lost their systems at the same moment, and each machine had to be repaired by hand. Neither the banks' own code nor an attacker was at fault: the risk arrived through an update from a supplier that everybody trusted.

Links to other zones · 16

Careers in this zone · 7

Core banking architect

Designs how a bank's systems fit together: accounts, cards, payments, accounting, the mobile app. Runs the technical side of replacing or modernising core banking — one of the riskiest projects in a bank's life.

What you need to know

  • The architecture of core banking and payment systems
  • Banking products and accounting (the general ledger)
  • Integrations: APIs, message queues, ISO 20022
  • Regulatory requirements on IT risk and outsourcing

Skills

  • Systems thinking
  • Selecting and managing vendors
  • Planning migrations
  • Explaining technical risk to the board

Certifications: TOGAF; core banking vendor certifications; cloud architecture certifications

Cybersecurity analyst at a financial institution

Monitors attacks on a bank or a payment system, investigates incidents, tests defences and prepares the team for a crisis. What is specific to finance: attacks on payment channels, fraud against client accounts, and tight deadlines for reporting incidents to the regulator.

What you need to know

  • Network security and running a security operations centre (SOC)
  • PCI DSS, the SWIFT Customer Security Programme, DORA
  • Typical attacks on the financial sector: phishing, ransomware, compromise of payment systems
  • Incident response

Skills

  • Log analysis, SIEM
  • Threat hunting
  • A cool head during an incident
  • Reports for the regulator

Certifications: CISSP, CISM, OSCP; GIAC certifications

Anti-fraud systems engineer (fraud analytics)

Builds the engine that scores every transaction in milliseconds — a card payment, a transfer, a login to the app — and decides whether to let it through or stop it. Develops features and models (device, geolocation, user behaviour), tunes the rules and makes sure the system catches fraudsters without blocking honest clients.

What you need to know

  • The architecture of real-time scoring systems
  • ML models for rare events (imbalanced data)
  • Card scheme rules, 3-D Secure, chargebacks
  • Behavioural biometrics and device analysis

Skills

  • Python, SQL, stream processing
  • Feature engineering
  • Testing rules and models for false positives
  • Working with investigation teams

Certifications: CFE (ACFE); anti-fraud platform vendor certifications depending on the profile

Data scientist / ML engineer in finance

Builds models for credit scoring, churn prediction, insurance pricing and document processing. The specific part: models must be explainable, non-discriminatory and independently validated — the regulator may ask why a client was refused a loan.

What you need to know

  • Statistics and machine learning
  • Credit risk and model metrics (Gini, stability)
  • Model risk management; the 🇪🇺 AI Act classifies credit scoring as a high-risk system (requirements from December 2027)
  • Data protection (GDPR)

Skills

  • Python, SQL
  • MLOps
  • Model interpretation (SHAP)
  • Communication with risk management and the regulator

Certifications: Not required; CQF, FRM or cloud ML certifications are a plus

RegTech engineer

Automates regulatory processes: client checks (KYC), sanctions screening, transaction monitoring, producing reports for the regulator. Translates the requirements of the law into rules and code that can be tested and explained to an auditor.

What you need to know

  • AML / KYC and sanctions requirements
  • Regulatory reporting formats (XBRL)
  • Data governance and data lineage
  • The basics of financial sector regulation

Skills

  • SQL, Python
  • Business analysis of requirements
  • Tuning rules against false positives
  • Documentation for audit

Certifications: CAMS; DAMA CDMP for data management

Site reliability engineer (SRE) / cloud engineer at a financial institution

Is responsible for keeping payments and applications running 24/7: builds fault-tolerant infrastructure in the cloud and in data centres, sets up monitoring, runs incidents and tests disaster recovery. In finance a few minutes of downtime means thousands of unprocessed payments and questions from the regulator.

What you need to know

  • Cloud platforms and containers
  • Fault tolerance and disaster recovery
  • Cloud and outsourcing requirements in the financial sector (🇪🇺 DORA)
  • Infrastructure security

Skills

  • Linux, Kubernetes, Terraform
  • Monitoring and observability
  • Incident management
  • Automation

Certifications: AWS / Azure / GCP; CKA (Kubernetes)

Blockchain developer / smart contract auditor

Writes or reviews smart contracts — programmes that execute financial deals on a blockchain by themselves: exchange, lending, token issuance. An error in the code here often means an irreversible loss of funds, which is why auditing code before launch is a profession of its own.

What you need to know

  • Ethereum / EVM, Solidity
  • Typical smart contract vulnerabilities
  • Applied cryptography
  • The mechanics of DeFi and tokenisation

Skills

  • Reading and testing code
  • Fuzzing and formal verification
  • On-chain transaction analysis
  • Fast response to attacks

Certifications: No formal standards — public audits, bug bounties and security contests are what count

All zones of the Atlas