Loading
Emerging technology
I do not collect technologies. I use them to solve problems that are already in front of me — in security, in civic infrastructure, and in the daily work of building software that survives real conditions. Beyond the screen, I am studying applied sciences, mechatronics, physical automation, drone technology, cyber-physical systems, cybernetics, and brain–computer interfaces — with one eye on post-von-Neumann computing.
Where the time goes
These are the areas I am actively building, practising, or testing. Each one is tied to public work you can inspect.
Systems where several models, tools, and retrieval layers cooperate on a real task instead of producing a demo. The interesting engineering is in the boundaries: what the agent may touch, how it recovers, and how a human audits the result.
Threat modelling as a design activity, not a compliance ritual. I build detection and remediation workflows that fit the way teams already ship, and I keep the Cisco CyberOps and ethical hacking material close to the code.
Computer vision is only useful when it changes a decision. ROADvision pairs pothole detection with satellite context and flood-erosion signals so that a prediction can become a maintenance queue rather than another dashboard. The repository documents five layers, including the institutional ones still in design.
The parts of the stack most teams inherit without questioning: offline behaviour, resilience under failure, and the ergonomics of the tools we use every day. These projects are where I test opinions I intend to keep.
The same systems discipline applied to public goods and cultural work. ROADvision treats road maintenance as a data problem; FEPAR treats African creative ownership as an infrastructure problem. Neither is a hackathon demo, and neither pretends the institutional layer already exists.
Efficiency is an engineering requirement, not a footnote. I design for lower compute, smaller payloads, longer device life, and measurable energy awareness — and I am public about the difference between a commitment and a verified carbon claim.
Applied science & physical systems
These are research and learning interests, not services I claim to have delivered. They are the direction of travel: sensing, control, motion, and the data pipeline that turns a physical event into a decision.
Physics, materials, and the applied science behind the tools we use. I follow the research and try to understand the mechanism, not just the API.
The seam where mechanical systems, electronics, and software meet. Motors, sensors, control loops, and the firmware that keeps them honest.
Software that moves something in the real world: conveyors, access systems, inspection rigs, and the safety logic that has to work every time.
Aerial sensing as a civic and industrial tool — survey, inspection, mapping, and the data pipeline that turns a flight into a decision.
The place where code stops being an abstraction and starts opening valves, moving vehicles, and switching grids. I study the safety cases, the OT/ICS security models, and the failure modes that only appear when software and the physical world share a control loop.
Feedback, control, and the boundary between people and machines. The older discipline underneath autonomy: how a system senses itself, corrects itself, and decides what counts as an error.
Direct signal paths between neural activity and computation — assistive technology first, augmentation second. Followed with a serious interest in the signal processing, the latency budget, and the consent and dignity questions that arrive before the hardware is ready.
Quantum computing, neuromorphic architectures, and the search for models of computation that are not merely faster but differently shaped. Studied seriously as a frontier, stated honestly as an interest.
Method
The list of interesting technologies is infinite. The list of problems worth solving is not. This is the filter I use.
A technology earns attention when it removes a constraint I can name: unreliable networks, an audit trail that has to exist, a model that must explain itself. Novelty alone is not a reason to learn something.
A working repository with a README beats a folder of slides. I would rather publish an unfinished system that runs than a polished claim that cannot be verified.
Every experiment is allowed to end. The point is to carry forward the pattern that worked, write it down, and use it in the next product instead of treating the stack as an identity.
Current experiments
A selection of the repositories behind this page. Each links to the source.
Iridescent DevSecOps Platform
An AI orchestrator for code review and vulnerability remediation.
Threat Detection Platform
Specialised agents that detect, analyse, and respond in real time.
ROADvision Lagos
An AI infrastructure sentinel for Nigerian roads.
ATS Résumé Matcher
High-precision candidate matching without the hype.
OchukoAi
A multi-agent framework for real enterprise work.
Rezilient.JS
An offline-first framework built for unreliable reality.
Online Banking Demo
A fintech interface built with a security-first lens.
iCryptex Exchange
A full-stack build around authenticated exchange records.
Ola Western
A client site built for clarity and speed.
Work together
I take on engineering roles, consulting engagements, and selected collaborations. The best conversations start with the constraint, not the technology.
