I’m a second-year Electrical Engineering student at Cape Peninsula University of Technology (CPUT) and an open-source developer/contributor at Wecncode. I am passionate about cybersecurity, artificial intelligence, and building tools that protect systems and users.

I am working towards a career in Security Operations Center (SOC) analysis, focusing on how machine learning, NLP threat intelligence, and application security guardrails mitigate zero-day social engineering and prompt injection threats.

• 🎓 Education: Studying Electrical Engineering at Cape Peninsula University of Technology (CPUT)
• 🚀 Open Source: Developer & Contributor at Wecncode
• 🛡️ Focus Areas: SOC Analysis, AI Threat Detection, Application Security & Incident Response
• 🤖 Specialization: LLM Security, NLP Social Engineering Detection & Deterministic Guardrails
• ⚽ Personal Interests: Swimming and Football

Languages & Databases

AI & Security Frameworks

Tools & Infrastructure

• Description: An open-source, AI-powered NLP tool designed to detect psychological manipulation, artificial urgency, and zero-day social engineering tactics in emails and chat messages before a victim is coerced.
• Key Focus: Semantic intent analysis, rule-based psychological heuristics, and zero-day BEC (Business Email Compromise) mitigation.
• Tech Stack: Python, NLP, Transformers, Pytest, Regex Semantics

• Description: Developed a local customer support AI agent with SQL tool execution to evaluate resistance against OWASP GenAI risks (LLM01 Prompt Injection & LLM08 Excessive Agency).
• Key Focus: Demonstrated that server-side ownership checks, parameterised SQL, and input validation reliably block unauthorized operations even when LLM prompt instructions fail.
• Tech Stack: Python, Ollama (llama3.2:3b), SQLite, Pytest

• Description: Transformer and Machine Learning pipelines built to intercept malicious user prompts (jailbreaks/instruction manipulation) and filter out email social engineering spam.
• Tech Stack: Python, Scikit-Learn, PyTorch, DeBERTa

• 🛡️ SOC Analytics & Threat Hunting: Log analysis, SIEM workflows, and MITRE ATT&CK Matrix.
• 🤖 AI Threat Modeling & Security: MITRE ATLAS framework, prompt guardrails, and sandboxing agent execution.
• 🔐 LLM Defense Engineering: XML tag sandboxing, zero-trust middleware for agentic tool use, and input sanitization.

I am actively seeking Cybersecurity Internships, SOC Analyst Graduate Roles, and Open-Source Collaborations in AI Security.

• GitHub: @Kamogelo0911
• Organization: @Wecncode
• LinkedIn: Kamogelo Ngwenya