AI Penetration Testing Specialist

International Turnkey Systems - ITS; Kuwait City; On-site; KWD 36,000 - 60,000; Full time

Job summary

Kuwait Islamic Bank in Kuwait City seeks an AI Penetration Tester to lead hands-on security testing for LLM/GenAI, ML, and AI-enabled applications, including chatbots and model endpoints. The role blends traditional PT with secure code review and vulnerability assessment focused on AI risks per OWASP Top 10 for LLMs.

Candidates should assess Python-based AI development and Azure AI / Azure OpenAI endpoints, design tests, build Python harnesses, and apply AI red-teaming methods using PyRIT and

Qualifications

• Bachelor's degree in computer science, Cybersecurity, Software Engineering, or related discipline.

Responsibilities

• Plan and execute AI penetration tests for LLM/GenAI and ML solutions across the bank (AI apps, AI gateways, model endpoints, orchestration layers, RAG pipelines, plugins/tools).
• Design test cases aligned to OWASP Top 10 for LLM Applications (e.g., prompt injection, insecure output handling, sensitive info disclosure, supply-chain risks).
• Perform adversarial testing using threat-informed techniques and references such as MITRE ATLAS (adversary tactics/techniques for AI systems).
• Validate security of agentic workflows (tool invocation, function calling, permissions, memory, connectors) and verify least privilege + authorization boundaries.

Skills

Python

Git

CI/CD integration

Education

Bachelor's degree in computer science or cybersecurity or related

3-7 years in penetration testing / application security / red teaming

Experience testing LLM/GenAI systems (prompt injection, jailbreaks, data exfiltration patterns, agent/tool abuse)

Strong Python for security automation and test harness development

Azure security concepts and exposure to Azure AI / Azure OpenAI / Foundry deployments

Strong report writing and remediation communication

Tools

Burp Suite

Kali

scanners

API tooling

PyRIT

Foundry

Job description

For a well-known Islamic Bank in Kuwait, we are hiring an AI Penetration Tester to join VMPT team and lead hands-on security testing for LLM/GenAI, ML, and AI-enabled applications (chatbots, RAG, AI agents, model APIs, and AI-assisted business workflows). The role blends traditional penetration testing + secure code review + vulnerability assessment with AI red teaming focused on modern AI risks such as prompt injection and insecure output handling as documented by OWASP's Top 10 for LLM Applications.

Candidate is expected to have experience in assessing Python-based AI development and Azure AI / Azure OpenAI / Microsoft Foundry-hosted model endpoints.

Key Responsibilities

AI / LLM penetration Testing & Red Teaming

• Plan and execute AI penetration tests for LLM/GenAI and ML solutions across the bank (AI apps, AI gateways, model endpoints, orchestration layers, RAG pipelines, plugins/tools).
• Design test cases aligned to OWASP Top 10 for LLM Applications (e.g., prompt injection, insecure output handling, sensitive info disclosure, supply-chain risks).
• Perform adversarial testing using threat-informed techniques and references such as MITRE ATLAS (adversary tactics/techniques for AI systems).
• Validate security of agentic workflows (tool invocation, function calling, permissions, memory, connectors) and verify least privilege + authorization boundaries.

Automation & Tooling

• Build Python-based test harnesses for repeatable AI security testing (fuzzing prompts, jailbreak suites, regression tests, evaluation scoring).
• Use and extend PyRIT (Python Risk Identification Tool) or equivalent frameworks to probe for jailbreaks/harms and automate adversarial scanning.
• Where applicable, leverage Microsoft Foundry capabilities such as the AI Red Teaming Agent (preview) that integrates PyRIT red teaming features.

Secure code review & AppSec

• Conduct secure code review (particularly Python) for AI application layers: prompt templates, RAG retrieval logic, output post-processing, input validation, secrets handling, authZ checks, and logging.
• Identify and help fix design flaws enabling data leakage, prompt manipulation, unsafe tool usage, or insecure output flows.

Traditional PT

• Perform web/API/cloud penetration testing for AI-enabled applications (auth, session, API abuse, SSRF, injection, misconfigurations, secrets).
• Run vulnerability assessments, triage findings, define severity, and validate remediation (re-test).

Education and Qualification

• Bachelor's degree in computer science, Cybersecurity, Software Engineering, or related discipline.
• 3-7 years in penetration testing / application security / red teaming.
• Demonstrated experience testing LLM/GenAI systems (prompt injection, jailbreaks, data exfiltration patterns, agent/tool abuse).
• Strong hands-on Python for security automation and test harness development.
• Practical experience with Azure security concepts (identity, networking, secrets, logging) and exposure to Azure AI / Azure OpenAI / Foundry style deployments.
• Strong report writing, clear communication, ability to work with developers on remediation.

A Must Certificates (at least one)

• OffSec OSCP / OSCP+
• GIAC GPEN
• Microsoft Certified: Azure Security Engineer Associate (AZ-500)
• ISC2 CCSP

Strongly Preferred

• OffSec OSWE (advanced web exploitation / white-box testing)
• Burp Suite Certified Practitioner (BSCP)
• Microsoft Certified: Azure AI Engineer Associate (AI-102) (helps align with Azure AI implementation patterns).

Tools & Technologies Skills

• Python, Git, CI/CD integration
• Burp Suite, common PT toolchains (Kali, scanners, API tooling)
• Azure security controls (identity, key management, logging/monitoring)
• AI red teaming tooling (e.g., PyRIT / Foundry scans)