Infrastructure
Cloud and on-premise architecture engineered for uptime and sized for growth.
We build the AI systems MENA institutions run on.
Applied AI, data infrastructure, and digital platforms for healthcare, government, and enterprise — designed in Cairo, delivered end to end, and measured in production.
- 120,000+ records digitized
- Bilingual by design — عربي · EN
TRUSTED ACROSS MENA
Egyptian Ministry of Health
Mansoura Ophthalmic Hospital
Egyptian Ministry of Health
Mansoura Ophthalmic Hospital
120,000+1
Patient records digitized
40+1
Government health facilities supported
5,000+2
Employees onboarded
60%1
Cut in admin overhead
1 Mansoura Ophthalmic Hospital and Egyptian Ministry of Health engagements, 2024–2026. 2 Retailo Academy LMS, MENA operations.
What we do
From architecture to operations
Six practices, one standard: systems that hold up in production, with the documentation to prove it.
Database architecture
Design, migration, and tuning of the data layer institutions depend on.
Digital transformation
Paper processes become digital workflows, with records that stay clean at scale.
AI analytics
Models and Arabic-capable NLP pipelines that turn operational data into decisions.
Interactive dashboards
Executive and operational views built for clarity in real time.
Decision support
Recommendation and risk-scoring systems that keep a human in charge.
How we work
From first call to steady state
A written, measurable process — so you always know what happens next.
1
Discover
We map your workflows, data, and constraints on the ground — not from a template.
2
Architect
You receive a written system design with scope, timeline, and success metrics before we build.
3
Build & integrate
Short cycles with working software every few weeks, integrated with your existing systems.
4
Operate & improve
We run, monitor, and tune the system in production — one accountable team.
Our research
حكيم
Hakeem-7B — an open Arabic medical reasoning model, built in Cairo
Open Arabic medical models sit at 40–50% on MedAraBench, below the level where output can be trusted. We trained a 7-billion-parameter model that reasons in Arabic and crossed the 60% line — then published the model and the method.
- DAPT · 14,404 documents
- SFT · 9,714 reasoning pairs
- DPO · 387 preference pairs
On a 14-model shared-parser leaderboard, Hakeem-7B ranks above every open model in the ≤8B class and above OpenBioLLM-70B and Llama-3.1-70B — statistically level with Gemma-2-27B and Med42-70B.
Source: MedAraBench clean split, think-first mode. Full ladder in the build log, 2026.
Our research
فيوبِرت
vioBERT-v3 — the first open Arabic medical BERT, built in Cairo
General Arabic encoders only see news and social text; English biomedical models never see Arabic. We domain-adapted MARBERTv2 on 1.12 million Arabic medical documents — cutting clinical perplexity 82.7% and lifting fill-mask accuracy 15.6 points — then published the encoder and the recipe.
- DAPT · 1.12M documents
- Whole-word masking · 22K steps
- Apache 2.0
vioBERT-v3 also cuts masked-LM perplexity 82.7% (24.34 → 4.21) and adds significant downstream classification gains (+1.62pp macro-F1, 95% CI above zero).
Source: 199 held-out masked items · DataCollator MLM, 15% masking, 2026.
Our research
أشعة
Ar-CXR — the first native Arabic chest-X-ray vision–language model
No open model could read a chest X-ray and write a report in Arabic. We coupled a frozen RAD-DINO vision encoder to a Falcon-H1-7B Arabic decoder — 7.4 billion parameters, only ~4% trained — that grounds 11 findings and writes a four-section report, ranking first on every Arabic metric.
- 223,462 CheXpert images
- RAD-DINO + Falcon-H1
- ~4% trainable
Ar-CXR grounds 11 findings at 0.789 macro-AUROC and leads METEOR, chrF, BERTScore and CIDEr — all under an identical 200-image, image-only Arabic protocol.
Source: clinical Jaccard on Arabic report generation, same prompt across all models, 2026.
In production
Platforms our clients run daily
Purpose-built systems for healthcare, education, and enterprise operations.
Healthcare · National eye hospital
Hospital Booking System
The public booking site and staff console running the front desk of Mansoura Ophthalmic Hospital. Patients book themselves across seven services with live, shared availability; staff manage scheduling, records, and reporting from a secure Arabic dashboard.
- Real-time scheduling with shared availability across 7 services
- Patient self-booking with national-ID checks and an instant printable confirmation
- Secure staff console with three access levels
- Live dashboards: daily and monthly load, service mix, no-shows
Live in 2 hospitals24/7 self-bookingArabic-firstNo-show tracking
Healthcare · Clinical documentation
Medical Report System
Clinicians capture the eye-exam once; the system turns it into a finished, tamper-proof Arabic medical report in one click — no retyping into Word. Every report is digitally signed and carries a scannable code that lets anyone confirm it is genuine.
- Guided clinical intake, then a one-click report with blank fields left out
- Tamper-proof: digitally signed with a scannable verification code
- Private, encrypted patient records with a full audit trail
- Keeps working without internet, then syncs; turnaround and doctor reporting
In productionOne-click reportsTamper-proofWorks offline
Enterprise learning · 5,000+ users
VionexLMS
An enterprise learning platform in production for more than 5,000 people. Build courses across ten content types, run monitored exams, deliver live instructor-led sessions, and issue certificates automatically — with separate administrator, instructor, and learner areas and a companion mobile app.
- Course builder — 10 content types (video, documents, exams, live sessions…)
- Monitored exams (camera/screen), a question bank, graded assignments
- Progress reporting: people, courses, teams, training matrix
- Branch and team scoping, one-click company sign-in, automatic certificates, mobile app
5,000+ learnersWeb & mobileAuto certificatesLive + on-demand
Client outcomes
Documented engagements, verifiable results
From national government agencies to regional enterprises — every outcome below is one our clients can verify.
Egyptian Ministry of Health
Centralized data infrastructure across government health facilities, with unified patient data management and reporting.
- 40+ facilities supported
- Unified national reporting
Mansoura Ophthalmic Hospital
Official AI & Innovation Hub: records digitization, intelligent booking, and automated workflows.
- 120,000+ records digitized
- 60% less admin overhead
Retailo Academy LMS: enterprise onboarding and training with real-time progress tracking across MENA operations.
- 5,000+ employees onboarded
- MENA-wide rollout
Vionex transformed our entire patient management workflow. The digitization of over 120,000 records and the intelligent booking system has fundamentally changed how we serve our patients. They didn't just deliver software — they understood our operational challenges.
The Retailo Academy LMS that Vionex built for us has onboarded over 5,000 employees across MENA. Real-time tracking, intuitive design, and rock-solid reliability. They delivered exactly what we needed, on time and beyond expectations.
PDF · 4 pages
Request the detailed healthcare case study
The Mansoura Ophthalmic Hospital engagement in depth: scope of work, our process, and the documented outcomes. The download starts as soon as you submit.
Who we are
A MENA-native AI house, built by practitioners
Vionex Digital Solutions is a technology company founded by AI practitioners with formal academic training and hands-on enterprise experience. We work the full distance between research and deployment: the same team that reads the papers ships the systems.
From national health infrastructure to enterprise analytics, our systems are designed for scale, documented outcomes, and Arabic as a first-class language rather than an afterthought.
-
Credentialed founders
Degrees in AI, with formal training in machine learning, NLP, and data science.
-
MENA-native, Arabic-fluent
Regional regulation, Arabic-language AI, and the operating realities of local institutions.
-
End-to-end ownership
Architecture, build, deployment, and ongoing operation — one accountable team.
From the blog
Notes from the build
Engineering decisions, research breakdowns, and field reports from teams shipping AI in healthcare, education, and government.
For procurement & IT
Questions enterprise clients ask us
Where is our data hosted?
Wherever your policy requires: public cloud in the region of your choice, your national cloud, or fully on-premise inside your own infrastructure. Government deployments frequently run entirely on-premise.
Do your systems work in Arabic?
Natively. Our platforms ship with bilingual Arabic/English interfaces, and our AI practice treats Arabic as a first-class language — we build and publish Arabic-language models.
Can you integrate with our existing systems?
Integration is the default, not an add-on. Our deployed platforms coordinate with existing records systems, notification channels, and reporting workflows rather than replacing everything at once.
How do you handle security and confidentiality?
Engagements begin under NDA. Production systems use role-based access control and audit trails, and deployments can be isolated entirely to your infrastructure when required.
Who owns the system and the data?
Your data is always yours. Ownership and licensing of deliverables are agreed clearly per engagement — before work begins, in writing.
What happens after launch?
We stay. Systems are operated, monitored, and improved in production under agreed service levels — the same team that built them.
Let's build your next system
Modernizing healthcare infrastructure, building enterprise analytics, digitizing operations — tell us what has to work, and we will scope it honestly.
Enterprise engagements begin under NDA. We reply by email.