Hire this AI Trainer
Sign in or create an account to invite AI Trainers to your job.
No software listed
Text
Document
Red Teaming
Entity (NER) ClassificationAI Red Teamer / LLM Security Engineer. Professional background includes roles such as AI Red Teamer / LLM Security Engineer and AI Security Researcher — Confused-Deputy IDOR in a ReAct Banking Agent. Education includes BSc, Kenyatta University and Certified AI/ML Pentester (C-AI/MLPen). AI-training focus includes data types such as Text and Document and labeling workflows including Red Teaming and Named Entity Recognition.
DocumentRed TeamingEntity (NER) ClassificationResearch write-up in progress demonstrating document-borne prompt injection through RAG-retrieved court filings against an agentic legal assistant, in Lakera Agent Breaker / Clause AI scenarios.
TextRed TeamingEntity (NER) ClassificationBuilt an autonomous reconnaissance-to-reporting pipeline as a 7-module Python system, with prompt sets tuned to the OWASP Web Top 10.
TextRed TeamingEntity (NER) ClassificationDiscovered and exploited a confused-deputy authorization flaw (IDOR) enabling cross-account data access through a tool-using ReAct agent; mapped to OWASP LLM06 (Excessive Agency) and MITRE ATLAS. Chained a second exploit via output-channel redirection and token-shattering to bypass input filtering. Published the technical write-up “Prompt Injection Against a ReAct Banking Agent.”
TextRed TeamingEntity (NER) ClassificationOffensive AI security practitioner specializing in adversarial testing of LLM applications and agentic systems. Applied a find-fix-prove methodology: discovered vulnerabilities, built exploit proofs of concept, and engineered or demonstrated mitigations. Work mapped to OWASP LLM Top 10, MITRE ATLAS, and NIST AI 100-2. Selected work included a confused-deputy IDOR in a ReAct banking agent, a deliberately vulnerable 3-agent LangGraph system with 10 documented vulnerabilities, a multi-agent bug-bounty pipeline, and research on indirect prompt injection via RAG poisoning. Competitive red-teaming included Gray Swan Arena, Lakera Gandalf, Microsoft AI Red Teaming Playground Labs, and Prompt Airlines CTF.
Continuous self-directed AI security program, AI Security
Certified AI/ML Pentester (C-AI/MLPen), AI/ML Penetration Testing
AI Red Teamer / LLM Security Engineer
AI Security Researcher — Confused-Deputy IDOR in a ReAct Banking Agent
