Proof of
excellence.

Problem

Security teams drown in disconnected findings. Scanners produce thousands of rows, but nothing tells you which weaknesses chain together into a realistic attack path — so remediation is prioritised by severity score instead of by actual exploitability.

Solution

I built an ingestion layer that accepts CSV security datasets and a prediction engine that processes structured JSON findings, correlating vulnerabilities, threats and security events against MITRE ATT&CK techniques, CVE records and CWE weakness classes. Automated risk-analysis workflows surface vulnerability relationships, lateral-movement risk and probable next-stage attacker activity.

Value

Defenders see correlated threats and predicted attack scenarios through interactive visualisations, with automated reporting that maps vulnerabilities to affected assets, techniques, risk levels and likely progression.

Problem

Assessment work was fragmented across separate tools with incompatible outputs, so every engagement ended in manual re-entry before any correlation or reporting could begin.

Solution

I built three dedicated modules — Network Security Testing, Web Security Testing and Endpoint Security Testing — behind one standardised JSON findings schema. The network module scans authorised environments for exposed services and weaknesses; the web module runs multiple attack scenarios against authorised applications; the endpoint module identifies misconfiguration and hardening gaps.

Value

Assessment results transfer automatically into the EDP prediction engine for ATT&CK correlation, with centralised visualisation of vulnerabilities, affected systems, techniques, risk relationships and remediation priority.

Problem

Training operations were running on scattered tools, making cohort administration, assessment and finance tracking inconsistent and impossible to scale.

Solution

I built a complete platform with a Student Portal and Instructor Portal covering online learning delivery, assessments, live classes, course administration and financial management for real training operations.

Value

Consistent, organised learning experiences that support self-paced study, instructor visibility into learner progress, and programme administration at scale.

Problem

Learners meet embedded and connected systems as theory, and most IoT prototypes are built with security bolted on afterwards — if at all.

Solution

I run a practical build track across Arduino, ESP32 and Raspberry Pi covering embedded systems, sensor integration, automation, smart devices, industrial IoT and educational robotics, with secure-by-design architecture patterns taught from the first prototype.

Value

Prototypes that work as teaching hardware and as the seed of future community solutions, with edge computing, basic computer vision and AI integration layered in as learners advance.

Problem

Local sellers had no simple way to reach online buyers or earn from products they do not stock themselves.

Solution

Founded and built a storefront covering lifestyle, fitness and fashion categories, with a dropshipping layer so partners can list and sell products for commission.

Value

A running retail brand with an income stream for resellers and a repeatable playbook for launching new product lines.

Problem

Talented Kenyan students and professionals lacked accessible, practical training paths into global technology careers.

Solution

Founded and developed Staken Hub Academy: a training institute with campuses in Malindi and Nairobi delivering hands-on technical programmes for high school graduates, university students and professionals upskilling.

Value

A growing pipeline of job-ready African tech talent, trained on practical, industry-aligned curricula.

Problem

Requisitions moved on paper, approvals stalled with no audit trail and departments could not see the status of their requests.

Solution

Built a role-based requisition workflow with department submission, multi-level approval, status tracking and an audit log, secured with row-level access rules.

Value

Approvals move in hours instead of days, and every request has a verifiable trail from submission to fulfilment.

Problem

Students had no self-service access to their records, so every status or result request went through staff manually.

Solution

Built an authenticated student portal exposing profile, enrolment, progress and records with strict per-user data isolation and an admin view for staff.

Value

Students self-serve their academic information and staff time shifts from record lookups to teaching.

Problem

Manually reviewing and grading programming assignments is time-consuming, inconsistent, and difficult to scale, especially when instructors need to evaluate large numbers of student GitHub repositories against detailed technical rubrics.

Solution

Built an automated code-grading system that integrates with GitHub repositories and uses AI to analyze student implementations against predefined rubrics. The platform automates repository retrieval, code analysis, rubric-based evaluation, scoring, and feedback generation.

Value

Reduces the time required to assess programming assignments while providing consistent, rubric-driven evaluations. It enables instructors to scale practical programming assessments and gives students actionable feedback on their implementations.