01Defensive cybersecurity · software engineering

I build secure systems
and prove it with numbers.

I’m Omar Alghamdi, a Saudi student in defensive cybersecurity. I build software people use every day, and I test it until the numbers hold.

02Work

Work#

Fourteen systems across 2025 and 2026, eight of them for clients. Each card opens the full story, with the numbers behind it.

9 screens

FinalistCyber-physical defense · 2026

SAQR

Every camera has to keep earning trust. Across 2,400 simulated runs, SAQR caught every modelled attack with zero false alerts.

2,400simulated runs100%detection

Read the case study saqrdefense.com
10 screens

Closed testingFamily safety · Google Play

Tatabbu

A family-safety app where a child’s location stays private by rules proven with SQL probes. It passed Google Play review in 14 minutes.

14 minPlay review176countries in closed testing

Read the case study No public link
9 screens

LiveRegulation as code · in production

Halqat Wasl

Used daily at a government school: a teacher’s note reaches the guardian only through the path the regulation defines.

22staff using it103students

Read the case study No public link
7 screens

Closed until launchPlatform with an autonomous guard

Alwaei Alfikri

The platform of the school awareness team I lead — guarded by an agent that screens every message, run at zero hosting cost, and closed until the school’s official launch.

60 sguard heartbeat3Durable Objects

Read the case study No public link
10 screens

LiveLearning platform · Moderneex

Dhura

A learning platform built on the Ministry textbooks, with interactive lessons for grades 5–9, 42,661 practice questions live and an AI that stays in the lesson.

46complete first-term courses, grades 5–91,340interactive lessons from the Ministry textbooks

Read the case study dhura.app
7 screens

LiveClinical reference · built with a pediatrician

PediaQuick

A pediatric reference in Arabic and English, built with a practising pediatrician: 143 clinical tools, every dose traced to its source and checked by a test.

143clinical tools: calculators, scores, scales and growth charts1,700+automated tests on the clinical core

Read the case study pediaquick.app

LiveClient delivery · freight

ProFleet

Carrier companies with their own drivers, sending offers to shippers — built into the client’s live app on the App Store and Google Play.

  • Join code per company
  • Offers, not price requests
  • Limits at two levels
  • Drivers migrated without loss
Short case No public link

LiveClient delivery · AI app

Smart Pen

The credits, payments, real-time voice and story studio behind a client’s AI app, live on the App Store, Google Play and the web.

  • Credit wallet with a full ledger
  • Receipts verified on the server
  • Live voice relayed by the server
  • One call per user, billed by the minute
Short case No public link
5 screens

LiveClient delivery · Android · iOS · Web

Beezato

Delivered on Android, iOS and the web with 476 automated tests, and checks that match every build to the contract.

476automated tests on one delivery3platforms: Android, iOS, web

Short case Google Play

LiveClient delivery · agriculture

AgriDustria

Each of the client’s 162 messages traced to a verdict with evidence, and the database hardened with a rollback for every change.

162client messages traced to a verdict with evidence

Short case No public link
7 screens

LiveClient delivery · education

Al-Warraq

A teacher picks the lessons and gets a print-ready exam on the school’s letterhead in about two minutes.

~2 minfrom lessons to a print-ready paper3paper types, each with its own typesetting

Short case al-waraq.com
6 screens

LiveClient delivery · real estate

Interactive 3D Property Map

Every unit boundary measured from the CAD plan to within 0.05% — not traced by eye over satellite images.

0.05%calibration error on both surveyed plot edges0.0–1.2%measured footprints vs declared areas

Competition entryKanz AI Training Hackathon · 2026

Dhaleel

A walking router that computes real shadows — up to about 68% less direct sun on the same walk. An AI Showcase Featured Project.

~68%less direct sun on the same journey, at best0servers — fully static

Short case No public link

In developmentArabic programming education · Moderneex

SeeScript

Teaches programming in Arabic: a Python curriculum of 9 units, 32 lessons, 179 auto-graded exercises and 4 projects. In development, not yet open.

  • Python curriculum
  • 179 auto-graded exercises
  • Arabic first
Short case No public link

03Proof

How the numbers were measured#

Every figure here comes from a logged run or a dated record, never an estimate. This is the method behind each one.

Five attack scenarios and one benign baseline, 100 runs each, against four systems — 2,400 runs in a digital twin.

Attacks detected — same attacks, four systems

Same attacks, same 2,400-run sweep, four systems. SAQR caught every modelled attack; each baseline missed some.

Evidence integrity
SAQR
each baseline

of evidence kept its integrity guarantee end to end

Evidence integrity, by system
Evidence integrity%
SAQR98
each baseline75
False alerts per hour
Vision-only 2.9
Network IDS only 3.8
SAQR 0
False alerts per hour, by system
False alerts per houran hour
Vision-only2.9
Network IDS only3.8
SAQR0

98% evidence integrity, against 75% for each of the three baselines. Zero false alerts, where the vision-only and network-only systems raised 2.9 and 3.8 an hour. Every number here comes from a run that was executed and logged, not from an estimate; the engine, the raw results and the security assessment are part of the project record.

Read the SAQR case study

04Labs

Try it#

Small working versions of the ideas behind the projects, running in your browser — illustrations of the published mechanism, not the engines themselves.

· SAQR

When a camera stops earning trust

Pick a camera and inject a freeze, a replay or a cut. Watch its trust fall, its state change and its evidence drop out of the decision.

The published rule
So the requirement became stricter: a camera that loses trust must stop being able to influence decisions at all. Not flagged — excluded. That single change moves the system from detection to containment, and it is much harder to build.

Illustration of the published mechanism, not the engine: the curve, the thresholds and the timings here are illustrative. Read the case study

Terminal

Terminal

An interactive terminal, simulated in your browser. Type a command, or press a chip below.

Or press / anywhere on the page.

05Recognition

Recognition#

Competition results, programmes and certificates — each one opens the document behind it — and the path that led here.

Honors#

SAIF 2026 FinalistSAQR · final, in-person stage · Riyadh, 24–26 Nov 2026

NSMO 2026 National FinalistInformatics · Mawhiba · after three qualifying stages · NSMO 2027: 100 / 100 at stage oneCertificate

AI Showcase Featured ProjectKanz AI Hackathon 2026 — DhaleelCertificate

Same programme

Participant in a Guinness World Records–setting AI lessonOne of 14,075 attendees · record achieved by Kanz with HRSD · 15 Jul 2026

Does not prove An individual title; the record belongs to the organisers.

Certificate

1,617 / 2,000Mawhiba Cognitive Abilities Test, 2024Certificate

Certificates#

Click any certificate to enlarge, or download it.

Competition4 stages2026

NSMO 2026 Informatics — through to the national finals

Mawhiba · Informatics

Proves Three qualifying stages passed — Cities & Governorates, Education Departments, National Teams — and a place in the National Finals of NSMO 2026, the National Science and Mathematics Olympiad run by Mawhiba with the Ministry of Education, Informatics track.

NSMO 2026 — Cities & Governorates
NSMO 2026 — Education Departments
NSMO 2026 — National Teams
NSMO 2026 — National Finals

Course2026

Harvard CS50P — Programming with Python

Proves Completing Harvard’s CS50P: nine problem sets and a final project (2026).

CS50P — Programming with PythonHarvard University · 2026

Assessment2024

Mawhiba Cognitive Abilities — 1,617 / 2,000

Proves A score of 1,617 out of 2,000 on the Mawhiba Cognitive Abilities Test (2024).

Cognitive Abilities — 1,617 / 2,000Mawhiba · 2024

Programme2026

AI Training Hackathon

AI Training Hackathon — Dhaleel featured

Kanz with LAU ACE · Jul 2026

Proves Completing the Kanz AI Training Hackathon (with LAU ACE, July 2026) and building Dhaleel, selected as an AI Showcase Featured Project — certificate ID KANZ-ADV-2183.

Case study: Dhaleel

Same programme

Guinness World Records event — participation

Participant in a Guinness World Records–setting AI lesson

One of 14,075 attendees · record achieved by Kanz with HRSD · 15 Jul 2026

Proves Attendance at the record-setting lesson.

Does not prove An individual title; the record belongs to the organisers.

Case study: Dhaleel

Course2024

Cybersecurity — Doroob

Proves Completing the Cybersecurity course on Doroob (Human Resources Development Fund, 2024).

Two fields are redacted — why

On the Doroob certificate two fields are covered with opaque boxes marked “Redacted”: the national identity number (in both languages) and the verification code at the foot of the page, because that code can be used to retrieve the number. Nothing else is altered, and the unredacted original is available on request.

CybersecurityDoroob · Human Resources Development Fund · 2024

Assessment2 tests

HackerRank — Python and Problem Solving (Basic)

HackerRank

Proves HackerRank’s basic-level certification tests, passed in Python and in problem solving.

Python (Basic)
Problem Solving (Basic)

Course4 levels2026

C++ track, 101–104 — Satr

Proves Completing the C++ course sequence on the Satr platform, from 101 to 104 (104 in Jun 2026).

C++ 101Satr platform
C++ 102Satr platform
C++ 103Satr platform
C++ 104Satr platform · Jun 2026

Course3 levels

Python track, 101–103 — Satr

Satr platform

Proves Completing the Python course sequence on the Satr platform, from 101 to 103.

Python 101
Python 102
Python 103

Programme2026

School appreciation — intellectual security programme

Proves A certificate of appreciation from Al-Rayyan Intermediate School (Taif) for outstanding effort in the school’s intellectual security programme and initiative — the Alwaei Alfikri team and platform (2026).

Case study: Alwaei Alfikri
Appreciation — intellectual security programmeAl-Rayyan Intermediate School · 2026
Download all certificates (.zip)

The path#

Every step aimed at one destination.

Mawhiba cognitive abilities test

Scored 1,617 out of 2,000 and entered the national gifted track.

Moderneex Studio founded

37 governance and policy documents, a brand identity, a 142-icon design system.

Fourteen systems built

Mobile apps on the App Store and Google Play, school platforms, a learning platform, a pediatric clinical reference, a 3D property map, a navigation agent and a cyber-physical defense system.

NSMO Informatics — through to the national finals

Cities & Governorates, Education Departments, National Teams, then the finals.

Kanz AI Hackathon — Dhaleel featured

An original AI project selected for the AI Showcase.

Participant in a Guinness World Records–setting AI lesson

One of 14,075 attendees · record achieved by Kanz with HRSD · 15 Jul 2026

Harvard CS50P completed

Nine problem sets and a final project, certificate issued.

Dhura goes live under Moderneex

A learning platform for grades 5–9: 46 first-term courses complete, with 1,340 lessons from the Ministry textbooks.

SAIF 2026 — SAQR through to the final stage

Final acceptance to the in-person exhibition in Riyadh, 24–26 November 2026.

NSMO 2027 — 100 / 100 at stage one

Informatics, in the Education Department competition test on 22 October.

Nabih — a play-to-learn game design, submitted

An AI-literacy game in which the player trains a robotic falcon with data, not code: a six-page game design document for the Ministry of Education’s play-to-learn competition (2026–2027), written with a classmate and now awaiting judging.

Governor, National Cybersecurity Authority

The destination. Everything above is preparation.

06About

About#

Who I am, how I work, and where I’m headed.

I’m Omar Hassan Alghamdi, a Saudi student who builds software real people use: apps on the App Store and Google Play, platforms that schools run every day, a pediatric clinical reference made with a practising pediatrician, and an interactive property map. I also built SAQR, a cyber-physical defense system for critical facilities — now a finalist at SAIF 2026 in Riyadh.

My field is defensive cybersecurity, and I treat it as a way of building rather than a title: access enforced in the data itself, claims proven by measurement, and checks run on the version that actually shipped.

My long-term goal is to lead Saudi Arabia’s national cyber defense as Governor of the National Cybersecurity Authority. Everything here is a step toward it.

How I work
  1. Prove, don’t assumeIf it hasn’t been measured, it isn’t known yet.
  2. Test what shippedChecks run on the version users actually get, not only on the source.
  3. Enforce access in the dataA screen can look right and still let the wrong person in.
  4. Fail safe, and say soWhen something goes wrong, the system stops trusting it and raises the alarm.
  5. Built to run for freeEverything I host myself runs at zero cost.

Skills#

No self-ratings: each skill links to the project or certificate where I used it.

07Contact

Contact#

For judges, mentors and security teams — WhatsApp and email reach me directly, and I reply myself. Open to mentorship and cybersecurity programmes.

Two CV formats on purpose. The designed sheet is for a human reader — English on the front, the same CV in Arabic on the back. The portal versions are single-column with standard headings, so an applicant tracking system reads them in the right order; PDF, DOCX and TXT hold the same text.

08How this page persuades

Persuasion has a research literature. This page follows it deliberately — and says so, because transparency is part of the credibility.

Each technique below is used on this page, and each is named here so that a reader can discount it.

  1. Credibility rests on two things: expertise and trustworthiness. So every claim here sits next to the method that verified it, not just its conclusion. Hovland & Weiss, Public Opinion Quarterly 1951; Pornpitakpan, J. Applied Social Psychology 2004
  2. Precise numbers read as confident and are believed more than round ones. So the figures are exact: 2,400 runs, 1,617 / 2,000, 14 minutes, 191.80 m. Jerez-Fernandez, Angulo & Oppenheimer, Psychological Science 2014
  3. The first number a reader sees pulls every judgment after it (anchoring). So each result is shown with its real baseline: 98% evidence integrity against 75% for each baseline system. Tversky & Kahneman, Science 1974; Janiszewski & Uy, Psychological Science 2008
  4. A signal is credible only when faking it would be costly. So the evidence is store releases, timestamped competition entries and test counts — things a claim cannot imitate. Spence, Job Market Signaling, QJE 1973
  5. Showing the work raises the perceived value of the result. So the method is on the page next to the number, and the audit runs in your own browser. Buell & Norton, Management Science 2011
  6. A story lowers counter-arguing and changes minds more durably than a list of claims. So each case study is told as a story: the real problem, the decision, the outcome. Green & Brock, JPSP 2000; van Laer et al., J. Consumer Research 2014 (meta-analysis); Green & Appel, Advances in Experimental Social Psychology 2024
  7. What is easy to process is judged more true. So sentences are short, one idea each, and every acronym is glossed on first mention. Schwarz, Jalbert, Noah & Zhang, Consumer Psychology Review 2021
  8. Concrete text is understood faster and remembered longer than abstract text. So it says “a function a stranger could still call”, not “robust access control”. Sadoski, Goetz & Rodriguez, J. Educational Psychology 2000
  9. A message that admits a weakness is believed more, not less. So there is a section titled Limits, and the four deep case studies each end with what the system still cannot do. Two-sided messages: Eisend, Int. J. Research in Marketing 2006 (meta-analysis); Crowley & Hoyer, J. Consumer Research 1994
  10. Competence without warmth reads as cold. So the teaching and community work is here too, not hidden. Stereotype Content Model: Fiske, Cuddy, Glick & Xu, JPSP 2002

No cookies, no analytics, no third-party requests; theme and language are remembered only in your browser.

omar@portfolio — zsh
Welcome. A terminal that runs in your browser — type help to begin, or press a chip.