Project map

Hover for details · click a marker to view

CompletedIn progressPlanned

Software Engineer · Data Science · Cybersecurity

Every project is another stamp in the passport.

I'm Rhulani Matiane, a software engineer focused on building robust applications and backend systems. I'm especially interested in data-driven systems and cybersecurity, using evidence to make software smarter, and secure engineering to make it worthy of trust.

View selected work About me →

Year applied experience

Selected projects

Core interests

Scroll to explore

About

Software engineering is the foundation

I build software first. Data science and cybersecurity expand how I investigate problems, measure decisions, and design systems I can trust.

Pretoria, ZA

I'm a software engineer and part-time Computer Science Honours student at the University of Pretoria. My foundation is software engineering: understanding the problem, designing the architecture, implementing the system, testing its behaviour, and leaving behind documentation another developer can actually follow.

Data science gives me another way to reason about software when a problem needs evidence, prediction, or measurement. Cybersecurity makes me think carefully about failure, access, misuse, and trust. Both make me a more deliberate engineer.

My work so far spans backend and platform design, explainable machine learning, digital forensics, multilingual NLP, and formal-verification research. The common thread is simple: readable code, measurable decisions, and systems people can understand after I have moved on.

Software architectureBackend developmentJava · Python · C++ · JavaScriptSQL · APIsData science & machine learningCybersecurityTesting & documentation

Engineering foundation

Build the system properly

I translate requirements into maintainable systems: clear boundaries, sensible data models, testable behaviour, and interfaces that another developer can extend.

Data & machine learning

Use evidence to improve decisions

I use data science to move beyond intuition: establish a baseline, choose meaningful metrics, evaluate limitations, and explain what the model can and cannot support.

Security mindset

Design for failure, misuse, and recovery

I treat validation, access, misuse, failure modes, and auditability as engineering concerns from the beginning rather than a checklist added at the end.

Current research

Turn complex evidence into useful guidance

My Honours research combines CBMC, ESBMC, and language models to turn formal-verification evidence into useful feedback for student C++ code.

Selected work

Projects that show how I think

A selection of work across software engineering, data science, and cybersecurity — focused on the problem, the technical decisions, and the evidence behind the result.

Completed · Software engineering

Mzansi Builds

Full Stack Web Application

A modern platform for builders to connect, collaborate, and build in public. Users can share progress, showcase projects through a live feed, and engage with other creators. The project emphasised intuitive UX, professional interface design, secure implementation, and maintainable software engineering practices. Its development followed an iterative SDLC and Agile workflow, supported by documented planning, testing, refinement, and version-controlled collaboration in the GitHub repository.

Full implementations

SDLC principles

Observed

System designUI/UX designFull stack developmentUnit, Integration and end-2-end testing

View GitHub repository →View live site ↗

Completed · Cybersecurity

ThreatSense

Explainable insider-threat detection

A hybrid Random Forest and XGBoost system for identifying risky employee behaviour. The engineering challenge was not only detecting anomalies, but producing an explainable risk score that an investigator could act on.

RF + XGB

Hybrid detection

Explainable

Risk scoring

PythonMachine learningSecurity analyticsModel explainability

View GitHub repository →

Completed · Cybersecurity/Software Engineering

API Threat Assessment Tool

Automated API security analysis

A comprehensive cybersecurity platform designed to automate the security testing of APIs, enabling organizations to identify vulnerabilities early, improve API resilience, and comply with industry security standards.

OWASP Top 10

Automatic scanning of OWASP Top 10 vulnerabilities

Executive and Technical Reporting

Comprehensive reporting

PythonWeb DevelopmentSecurity analyticsAutomation

View GitHub repository →

Completed · Data science

Cross-language AI text detection

isiZulu · isiXhosa · English

A multilingual detector evaluated across isiZulu, isiXhosa, and English. The project explores what changes when an NLP problem is designed for underrepresented languages instead of treating English performance as universal.

Languages evaluated

Test cases

PythonNLPMultilingual evaluationResponsible AI

View GitHub Repository →

In progress · Honours research

Formal verification + LLM feedback

Formal methods made useful to learners

An Honours research system that translates bounded-model-checker output into clear feedback for student C++ code. The goal is to make rigorous verification evidence useful without hiding or weakening it.

CBMC + ESBMC

Verification evidence

LLM

Pedagogical feedback

C++Formal verificationLLMsResearch

Ask about the research →

In development

What I'm building next

These are planned portfolio projects chosen to deepen my experience in developer tooling, automation, and production-minded data engineering.

Planned · Developer tooling

Developer automation toolkit

Python · APIs · workflow automation

Production-minded tools for report generation, document processing, API integration, and repetitive developer workflows designed with configuration, logging, and failure handling from the start.

PythonAPIsAutomationObservability

Discuss the roadmap →

Planned · Data engineering

Observable data pipeline

Ingestion · transformation · monitoring

A reliable data platform that validates incoming data, applies reproducible transformations, exposes useful outputs, and makes failures visible. The emphasis is reliability and observability, not just moving rows from A to B.

ETLData validationMonitoringReproducibility

Discuss the roadmap →

How I work

From problem to production-minded solution

My process is straightforward: understand the constraints, design the system, build an end-to-end solution, and improve it with tests, feedback, and evidence.

Understand the problem

Understand the user, constraints, risks, and success criteria before choosing the technology or writing the first line.

Design the system

Choose clear boundaries, data flows, and interfaces that solve the current problem without making the next change unnecessarily hard.

Build and validate

Get an end-to-end version working early, then refine it with tests, feedback, profiling, and measured evidence.

Document and improve

Validate assumptions, inspect failure modes, document the trade-offs, and leave the system understandable for the next engineer.

Engineering principles

How I want my work to hold up

Different projects demand different tools, but these principles stay consistent across the software I build.

Clarity over cleverness

Readable code, explicit assumptions, and names that reveal intent. A system should still make sense when its original author is no longer there to explain it.

Evidence over assumptions

Tests, profiles, metrics, and data should shape technical decisions. I would rather report a modest result honestly than defend an impressive guess.

Security by design

Validation, access control, misuse, recovery, and auditability are engineering concerns from the beginning, not paperwork added at the end.

Make systems explainable

Whether the output comes from an API, a model, or a verification tool, people should understand what produced it and how to act on it.

Let's build something worth shipping.

I'm looking for software engineering and developer opportunities where I can contribute to real products, grow alongside strong engineers, and bring an informed interest in data and cybersecurity.

Start a conversation →