AI Security Engineer

About the Role

We are seeking an experienced AI Security Engineer to join our team in Riyadh, Saudi Arabia. The ideal candidate will specialize in protecting AI-driven applications, including Large Language Models (LLMs), AI agents, Retrieval-Augmented Generation (RAG) pipelines, prompt workflows, and enterprise AI platforms. This role is pivotal in identifying and mitigating AI-specific security vulnerabilities such as prompt injection attacks, jailbreak exploits, data leakage, insecure tool usage, unsafe AI outputs, and misuse of AI systems.

The AI Security Engineer will collaborate closely with AI engineering, platform, cloud, and security teams to develop secured AI solutions, establish governance frameworks, and uphold responsible, compliant use of AI technologies.

Key Responsibilities

• Develop and enforce security measures for AI applications, LLM integrations, AI agents, and systems utilizing RAG methodologies.
• Detect and counteract AI-centered threats including prompt injections, jailbreaks, data exfiltration, model misuse, hallucination hazards, and unsafe tool/function calls.
• Protect prompt workflows, system prompts, agent instructions, retrieval pipelines, and AI orchestration layers from vulnerabilities.
• Evaluate risk factors related to LLM APIs, vector databases, embeddings, knowledge bases, and third-party tool integrations.
• Manage secure processing of sensitive data such as user inputs, retrieved contextual information, prompts, and AI model outputs.
• Perform threat modeling exercises for AI-based solutions, agents, and enterprise AI assistants.
• Set and monitor AI security governance policies alongside responsible AI controls and compliance standards.
• Carry out security assessments, red team activities, and system audits targeting AI infrastructures.
• Keep abreast of emerging AI-related security issues and propose proactive mitigation approaches.
• Partner with engineering teams to harden AI APIs, cloud deployments, access controls, and production AI platforms from security threats.

Qualifications and Requirements

• Over 5 years of practical experience with AI applications, including LLMs, RAG frameworks, embeddings, vector search technologies, prompt engineering, and AI agent deployment.
• Proven ability to identify and mitigate LLM security challenges such as prompt injections, jailbreaks, leakage of sensitive data, and improper output handling.
• Comprehensive knowledge of designing secure AI architectures encompassing authentication, authorization, logging, monitoring, and data protection mechanisms.
• Familiarity with cloud environments like Azure, AWS, or Google Cloud Platform, focusing on AI service implementation and API deployment.
• Experience utilizing security tools including Static and Dynamic Application Security Testing (SAST/DAST), Security Information and Event Management (SIEM), vulnerability scanners, and security monitoring solutions.
• Understanding of data privacy laws, regulatory compliance, and responsible AI usage standards.
• Hands-on experience with AI security frameworks, risk management processes, or governance related to responsible AI usage.
• Exposure to techniques such as LLM red teaming, establishment of AI guardrails, content safety measures, prompt filtering, and output validation.
• Expertise in securing RAG pipelines, AI agents, tool-calling mechanisms, and enterprise chatbot platforms.
• Awareness of OWASP Top 10 risks for LLM applications or similar AI security guidelines.
• Relevant certifications such as CISSP, CEH, Security+, or cloud security qualifications.
• Experience working within regulated domains such as telecommunications, fintech, or enterprise sectors.
• Strong analytical skills and problem-solving capabilities.
• Ability to collaborate effectively with cross-functional teams including AI, security, cloud, and product stakeholders.
• A proactive approach toward early detection of AI security risks prior to production deployment.
• Excellent communication skills to articulate risks and mitigation strategies to diverse audiences, including technical and non-technical stakeholders.