AI Security Engineer
Position Title - AI Security Engineer Education - Bachelor’s degree in computer science, Cybersecurity, Information Technology, Information Systems, Engineering or equivalent. - Postgraduate qualification in information/cyber security is an advantage. Minimum Work Experience - 5+ years of cybersecurity engineering experience, including cloud security, application security or data security, with practical exposure to AI/ML or agentic AI platforms. Skills / Certifications - Microsoft Azure security/AI certifications such as SC-100, AZ-500, AI-102 or equivalent - CISSP, CCSP or CISM (preferred) - AI security or ML security training; OWASP LLM knowledge - CKS/Kubernetes or DevSecOps certification (advantage) Preferred Skill Set - Azure AI/OpenAI/ML security and cloud-native controls - IAM, managed identities, API security, secrets and network isolation - AI threat modelling, prompt injection testing and secure RAG patterns - DSPM, DLP, classification and regulated data protection - Python/PowerShell, CI/CD and security automation Job Purpose Engineer and validate security controls for AI/ML and agentic AI services used by the client, including Azure AI Services, Azure OpenAI, Azure Machine Learning, AI-enabled applications and third-party AI platforms. The role protects sensitive and regulated data, identities, models, APIs, integrations and AI infrastructure throughout the AI lifecycle. Primary Responsibilities:- - Maintain an inventory of AI/ML assets, models, inference endpoints, agents, connectors, plugins, training pipelines and supporting infrastructure. - Perform security reviews for Azure OpenAI, Azure AI Foundry/Services, Azure Machine Learning and approved third-party AI services. - Assess authentication, authorization, managed identities, service principals, API keys, secrets, network exposure and rate-limiting controls. - Implement and validate least-privilege RBAC, private endpoints, firewall rules, network isolation, encryption and secure secret storage. - Assess AI data flows for PII/PHI and confidential data exposure across prompts, outputs, RAG repositories, connectors and training datasets. - Configure or validate content safety filters, prompt guardrails, custom blocklists and abuse-prevention controls. - Test for prompt injection, jailbreaks, insecure output handling, excessive agency, data leakage, model misuse and integration abuse. - Review model registry permissions, model/version integrity, dependency provenance and AI supply-chain security. - Develop AI-specific logging, monitoring and detection use cases for anomalous API activity, prompt attacks, privilege misuse and data exposure. - Support Shadow AI discovery and enforcement using approved CASB/SWG, endpoint and cloud controls. - Conduct threat modelling and security testing for AI use cases before production release and following material changes. - Track AI security findings, risks, exceptions and remediation actions to closure; support risk acceptance where required. - Produce AI security assessment reports, data exposure findings, control validation evidence and maturity recommendations. - Coordinate with AI Governance, Privacy, Legal, Data, Cloud, AppSec, SOC and business owners.