Model card

In plain English: who I am, what I do now, and what I want next, in three short paragraphs.

Card authored by the model. Conflict of interest: total. To compensate, every number below points to an exhibit you can check.

In plain English: the basic facts: name, job, location, and the doctorate that is still in progress.

In plain English: the kinds of jobs I am a good fit for, each with the proof.

Also accepts: research collaborations, and Cyntraix engagements: security assessments, Zero Trust architecture, risk governance and AI security risk.

In plain English: the jobs and claims I am not the right fit for, stated up front so nobody is surprised.

The model should not be used for the following. Listed here so nobody finds out in the interview.

In plain English: my education and every job I have held, in order.

Two corpora, collected over fifteen years. Overlaps are real; some years had two jobs.

Formal education

Work experience

Rows marked with a bar are still held.

Supplementary data

CompTIA Security+ (2026-05). Google Cybersecurity Professional Certificate (2023-11). App Academy AI-Powered Software Development and Generative AI Engineering (2025-07). Introduction to Generative AI, Google Cloud (2023-05). IC3 and a CMMC 2.0 course (2026-02). SANS AI Cybersecurity Forum (2024). Agentic AI Bootcamp, Howard AI Network (2026). Full ledger

In plain English: measured results, each taken from something real, with the setting it was measured in.

No benchmark was invented for this card. Each result is a count or a measurement from a real artifact, next to the setting it was measured in. Where the setting is a simulation or a home lab, the table says so.

In plain English: where my evidence is thinner than it looks, stated honestly.

• Small lab, labeled as such. The lab is two small machines at home, not an enterprise. Its numbers are real and small.
• Simulated network. Breakwater runs against simulated containers. Nothing in it is presented as client work.
• Preprints are not peer review. Twenty-four SSRN preprints show output, not refereeing. The ACM paper is still in preparation.
• Known gaps. No CI security pipeline, Infrastructure as Code, hands-on AWS or paid SOC work on file yet. Each has a small public repo planned, and none moves onto this card until it exists.
• Evidence bias. Fine-tuned on forensics, the model asks “how do we know?” before agreeing to anything. This is intended behavior.
• Author bias. The card was written by the model. Every number links to its exhibit for that reason.

In plain English: how I keep my own systems, and this website, safe.

Practices the maintainer applies to his own systems, including this page.

In plain English: my career as a dated list, newest first.

In plain English: how to reference this page in a paper or report.

If this model is useful in your work, please cite it. Better yet, hire it.

@misc{ologunde2026,
  author       = {Ologunde, Ezekiel},
  title        = {ezekiel-ologunde: a model card},
  year         = {2026},
  version      = {v2026.09},
  note         = {D.Eng. checkpoint expected 2027-09. ORCID 0009-0005-7207-8506},
  howpublished = {\url{https://ezekielologunde.github.io/}}
}

In plain English: short explanations of the technical terms used on this site.

Adversarial machine learning

Attacks that trick or corrupt AI models, and the defenses against those attacks.

AI

Artificial intelligence: software that learns patterns from data and makes suggestions or predictions.

Attack graph

A map of the steps an attacker could take, from a starting point to a target.

Capture the flag

A security exercise where students solve hacking puzzles to find hidden answers, called flags.

CI

Continuous integration: automated checks that run every time the code changes.

CMMC

Cybersecurity Maturity Model Certification: the U.S. Defense Department’s security standard for its contractors.

CPE

A standard name for a product and version, used to match it to its known flaws.

CVE

Common Vulnerabilities and Exposures: the public ID number given to each known security flaw.

CVSS

A 0 to 10 score of how severe a security flaw could be.

D.Eng.

Doctor of Engineering, a doctoral degree focused on applied research. His is expected in September 2027.

Differential privacy

A mathematical way to learn from data without exposing any single person’s or site’s records.

Digital forensics

Investigating computers and networks to find out what happened, and proving it with evidence.

EDR

Endpoint detection and response: software on each computer that watches for attacks and helps stop them.

EPSS

A 0 to 1 estimate of how likely a security flaw is to be used by attackers soon.

Ethical hacking

Hacking with permission, to find and fix weaknesses.

Explainable AI

AI whose decisions can be explained to a person, so they can be checked.

Federated learning

Training one AI model across several sites without sending their raw data to one place.

GRC

Governance, risk and compliance: making sure an organization manages its risks and follows the rules it must follow.

Host-based detection

Software on each computer that watches for signs of an attack on that machine.

IoT

Internet of Things: everyday devices on a network, such as cameras, speakers and thermostats.

KEV

The U.S. government’s (CISA) catalog of flaws that attackers are known to be using.

Language model

An AI system that reads and writes text, such as a chat assistant.

LLM

Large language model: the kind of AI behind chat assistants, which reads and writes text.

MQTT

A lightweight messaging system many smart devices use to talk to each other.

NIST

The U.S. National Institute of Standards and Technology, which publishes widely used security frameworks.

NVD

The U.S. National Vulnerability Database, the public record of known security flaws.

OT

Operational technology: computers that run physical equipment such as pumps, valves and production lines.

Penetration testing

Authorized, simulated attacks that find weaknesses before criminals do.

Preprint

A research paper shared publicly before, or without, formal peer review.

Prompt injection

Hidden instructions slipped into text an AI reads, trying to make it do something it should not.

Proxmox VE

Software that runs many virtual computers on one physical machine.

Reinforcement learning

Training an AI by trial and reward, the way a game-playing AI learns.

RLS

Database rules that decide which rows of data each user is allowed to see.

SBOM

Software bill of materials: a list of every component inside a piece of software.

SCADA

Supervisory control and data acquisition: systems that monitor and control industrial equipment from a distance.

SIEM

Security information and event management: software that gathers activity records from many systems in one place and raises alerts.

SSRN

An online library where researchers post papers, often before or without formal peer review.

Threat model

A structured list of what could go wrong in a system, how, and what stops it.

Zero Trust

A security approach where no user or device is trusted by default, even inside the network.

In plain English: how to reach me.

Write about a security engineering, AI security, GRC, detection or teaching role, a research collaboration, or a Cyntraix engagement.

ologundeomotola@gmail.com

• LinkedIn in/ezekiel-ologunde
• Google Scholar profile
• ORCID 0009-0005-7207-8506
• GitHub ezekielologunde
• Medium @ezekielologunde