The Architecture of Care: How I Built a Career at the Intersection of Healthcare Data and Human Impact

I didn’t set out to become a specialist in NHS cancer data pipelines. Nobody does. But somewhere between my first encounter with a broken HL7 message and the moment I watched a real-time oncology dashboard light up with data that had never existed in structured form before, I understood that healthcare technology isn’t just a career — it’s a responsibility.

This is the story of how a decade of obsessive focus on healthcare data architecture led me to build systems that matter, and why I believe the people who truly understand this space carry a rare and underappreciated kind of expertise.

Who I Am

My name is Ayoub Bensakhria. I am a Lead Software Engineer at ARO Technology in Liverpool, where I work on NHS projects that sit at the critical intersection of data engineering, interoperability, and clinical intelligence. I hold dual MSc degrees — one in Artificial Intelligence and one in Financial Engineering — and I have spent the last ten years specialising almost exclusively in healthcare technology.

I am also the founder and sole director of Yoctobe Ltd, a Liverpool-based software company I established in February 2025, building Laboratory Information System (LIS) and Medical Reporting System (MRS) platforms for healthcare providers across Africa. Yoctobe is not a side project. It is the long-form expression of everything I have learned about healthcare data — applied to a continent where the infrastructure gap is widest and the impact potential is highest.

I describe my USP in a single phrase: the art of architecture and a deep understanding of healthcare data.

A Decade Inside the NHS Data Engine

Most engineers know what an API is. Fewer know what a COSD submission looks like at 2am when a validation failure is blocking a cancer registry feed. I know both.

NHS Cancer Data: Building the Pipelines That Power Clinical Intelligence

My most significant technical contributions at ARO Technology have been in the ingestion, transformation, and submission of NHS cancer datasets. This is not glamorous work. It is precise, high-stakes, and unforgiving — and it is exactly the kind of work I find most meaningful.

The datasets I have worked with include:

• COSD (Cancer Outcomes and Services Dataset) — the backbone of NHS England’s cancer intelligence, capturing patient pathways from diagnosis through treatment and beyond. I have built ingestion pipelines that parse, validate, and submit COSD data to national registries, navigating the schema complexity and submission rules that catch out engineers who underestimate this domain.
• RTDS (Radiotherapy Dataset) — capturing radiotherapy treatment episodes at a granular level. Working with RTDS requires an understanding of both clinical workflows and the technical constraints of the submission infrastructure.
• SACT (Systemic Anti-Cancer Therapy dataset) — chemotherapy and systemic treatment records. SACT data sits at the intersection of pharmacy, oncology, and patient safety, making data quality failures more than a technical problem.
• National Registries — contributing to the broader landscape of cancer intelligence by ensuring data reaches the national bodies that use it to shape policy, resource allocation, and research.

What I have built in this space is not just pipelines. It is trust — trust that the data flowing through these systems is accurate, timely, and clinically meaningful.

NHS Terminology and Vocabulary: Speaking the Language of Clinical Data

Healthcare data without the right vocabulary is noise. I have developed deep fluency in the terminological standards that give NHS data its clinical meaning:

• SNOMED CT — the world’s most comprehensive clinical terminology, used across NHS systems to encode diagnoses, procedures, and observations in a way that is both human-readable and machine-processable.
• OPCS-4 — the classification system for surgical and procedural coding in UK secondary care.
• ICD-10 — the international classification of diseases, used for diagnosis coding and underpinning commissioning, reporting, and research across the NHS.
• dm+d (Dictionary of Medicines and Devices) — the NHS reference standard for medicines and devices, critical for SACT and pharmacy data workflows.

Understanding these vocabularies at depth is not optional when you are building systems that touch patient records. A wrong code is not a bug — it is a clinical error with real consequences.

FHIR-Based Interoperability: The Present and Future of NHS Data Exchange

The NHS is mid-transition toward HL7 FHIR (Fast Healthcare Interoperability Resources) as the standard for clinical data exchange, and I have been building in this space for years.

FHIR is elegant in theory. In practice, implementing it requires understanding not just the specification, but the NHS-specific profiles, extensions, and implementation guides that constrain it for the UK context — including NHS England’s national FHIR API standards and the UK Core FHIR profiles maintained by HL7 UK.

I have worked with FHIR R4 across multiple integration patterns: RESTful APIs, FHIR messaging, and document exchange. I understand the gap between what the FHIR specification says and what a live NHS system actually does — and I know how to bridge it.

OMOP and the Global Standard for Real-World Evidence

Beyond the NHS-specific landscape, I have worked with the OMOP Common Data Model (CDM) — the international standard for harmonising observational health data to enable large-scale analytics and real-world evidence generation.

OMOP is the foundation of the OHDSI (Observational Health Data Sciences and Informatics) network, used by research institutions, pharmaceutical companies, and health systems globally. Mapping NHS data to OMOP is a non-trivial exercise that requires both clinical domain knowledge and engineering rigour. I have done this work, and I understand the vocabulary mapping challenges, the ETL complexity, and the analytic possibilities it unlocks.

Interoperability Across Borders: UK and EU

Healthcare data does not stop at national boundaries — and neither does my experience.

In the UK, interoperability is increasingly driven by the NHS Spine, GP Connect, National Record Locator, and the developing Federated Data Platform. I have worked within this ecosystem, understanding how data moves between primary care, secondary care, and national systems.

In the European context, interoperability is shaped by the European Health Data Space (EHDS) regulation — a landmark EU framework for cross-border health data sharing that came into force in 2024. Through Yoctobe’s work serving African healthcare providers, I have also engaged with interoperability challenges in lower-resource settings, where the gap between international standards and on-the-ground realities demands creative, pragmatic engineering.

Big Data: Ingestion, Transformation, and the Engineering of Scale

NHS data is not small. Trusts generate millions of records annually, and national datasets aggregate across hundreds of organisations. I have built and maintained the ETL (Extract, Transform, Load) pipelines that move this data reliably at scale.

My engineering approach is shaped by several principles:

Data quality is a clinical responsibility. Every transformation decision has downstream consequences for analytics, reporting, and — ultimately — patient care.

Schemas change. Systems don’t wait. NHS datasets are versioned, and major schema updates (like COSD v9 and v10 migrations) require careful version management, backward compatibility planning, and close coordination with clinical informatics teams.

Volume is the easy problem. Provenance is hard. Knowing where data came from, what transformations it has undergone, and whether it can be trusted for a given use case — that is where real data engineering expertise lives.

Data Visualisation: Making the Invisible Visible

Data that cannot be understood cannot drive decisions. I have designed and built dashboards and visualisation layers that turn complex NHS datasets into clinical intelligence — from cancer pathway performance dashboards to radiotherapy workload analytics.

The challenge in healthcare visualisation is not aesthetics. It is clinical fidelity: ensuring that what a clinician or commissioner sees on a screen accurately reflects the underlying data, with appropriate context, caveats, and drill-down capability.

Regulation and Data Governance: The Framework That Makes Trust Possible

Healthcare data is among the most sensitive data that exists. The frameworks that govern its use are not bureaucratic obstacles — they are the architecture of trust.

I work within and across:

• UK GDPR and the Data Protection Act 2018 — the legal foundation for all NHS data processing.
• NHS Data Security and Protection Toolkit (DSPT) — the assurance framework that every NHS-connected organisation must comply with.
• Data Security Standards (DSS) — the ten standards that operationalise data security across NHS organisations.
• DTAC (Digital Technology Assessment Criteria) — the NHS’s framework for assessing digital health products before they enter clinical pathways, encompassing clinical safety, data protection, technical assurance, and interoperability.
• EU MDR (Medical Device Regulation) — relevant to Yoctobe’s SaMD (Software as a Medical Device) development, where our LIS and MRS platforms must meet the regulatory bar for clinical-grade software in African markets that reference European standards.

Governance is not a checklist. It is an engineering discipline — and treating it as such is one of the things that distinguishes serious healthcare technologists from those who are still learning.

Yoctobe Ltd: Taking Healthcare Architecture to Africa

In February 2025, I founded Yoctobe Ltd — a Liverpool-based SaMD and AI-AMD (AI-enabled Medical Device) development company with an entirely African client base.

Yoctobe builds two core platforms:

• LIS (Laboratory Information System) — enabling laboratories across Africa to digitise, standardise, and connect their workflows, from sample tracking to result reporting, with interoperability at the core.
• MRS (Medical Reporting System) — supporting medical reporting workflows with structured data capture, FHIR-aligned data models, and integration capabilities that allow African health systems to progressively connect with global health data standards.

The work is technically demanding. Building clinical-grade software for environments with variable connectivity, limited IT infrastructure, and evolving regulatory landscapes requires a different kind of engineering judgement — one that prioritises resilience, simplicity, and real-world usability alongside standards compliance.

Yoctobe is the product of ten years of NHS-grade engineering rigour applied to a context where the need is acute and the margin for error is just as real.

The Art of Architecture

I use the phrase art of architecture deliberately. Architecture in healthcare technology is not just about selecting the right database or designing an efficient API. It is about holding in mind, simultaneously:

• The clinical workflow that generates the data
• The regulatory framework that governs its use
• The technical standards that enable its exchange
• The human being whose care depends on its accuracy
• The system that will consume it in ways you cannot fully anticipate

Getting this right requires deep domain knowledge, strong engineering fundamentals, and something harder to name — a kind of structural intuition that comes from having built systems that failed in production, navigated schema migrations under deadline, and explained data quality issues to clinical informaticists who needed answers in plain English.

That is what I bring. Not just the ability to build — but the understanding of why it should be built a particular way, and what the consequences are if it isn’t.

Connect With Me

I write and think publicly about NHS data architecture, healthcare interoperability, FHIR implementation, and the technical realities of building clinical-grade software — both in the NHS and in emerging health systems.

If you are working on:

• NHS data integration or national dataset submissions
• FHIR implementation projects in UK or EU healthcare settings
• OMOP CDM adoption for research or real-world evidence
• Healthcare technology for African health systems
• SaMD regulatory strategy under DTAC or EU MDR

I am open to conversations, collaborations, and consulting engagements.