AI Security Specialist

Profectus; Canberra; Hybrid; AUD 180,000 - 240,000; Full time

Job summary

Profectus seeks an experienced AI Security Specialist to securely enable AI adoption in an Australian Government context. You will review AI-enabled systems, test for misuse, and advise teams to progress through assessment and authorisation.

You will conduct AI red teaming, design guardrails, and prepare assurance artefacts. Strong experience securing LLMs and RAG, along with ISM knowledge and NV1 clearance, is essential.

Qualifications

• Hands-on experience securing AI/ML systems, incl. LLMs and RAG.
• Knowledge of AI-specific threats and controls such as prompt injection, data leakage and guardrail design.
• Experience in Defence or Australian Government security environments.
• Ability to translate technical findings into clear risk advice for technical and non-technical stakeholders.
• NV1 security clearance or ability to obtain it.

Responsibilities

• Conduct security reviews of AI-enabled systems across models, prompts, data flows, hosting and supply chain.
• Perform AI red teaming and adversarial testing to identify weaknesses before deployment.
• Test LLMs, RAG and generative AI apps for prompt injection, jailbreaking, data poisoning and leakage.
• Assess security of tool/agent integrations within AI systems and recommend guardrails.
• Provide security evidence and input for assessment leading to Authority to Operate.

Skills

AI security

Security testing

Adversarial testing

Risk communication

Project coordination

Assessment & Authorisation

Azure/AWS security

ISMS/ISM knowledge

Education

Tertiary qualification in cyber security / CS / data science / engineering

Cyber security certifications

Tools

Azure AI

AWS AI/ML

Job description

Profectus is seeking an experienced AI Security Specialist to support the secure adoption of artificial intelligence within an Australian Government agency. You will examine AI-enabled systems for security weaknesses, test how models behave under misuse and attack, and advise project teams and risk owners so projects can progress through assessment and authorisation. The work covers security reviews, AI red teaming, control and guardrail recommendations, and AI assurance artefacts.

Key Responsibilities

• Conduct security reviews of AI-enabled systems, covering models, prompts, data flows, hosting, integrations and supply chain
• Perform AI red teaming and adversarial testing to identify weaknesses before systems are deployed
• Test LLM, RAG and generative AI applications for prompt injection, jailbreaking, data poisoning and model and data leakage
• Assess the security of tool and agent integrations within AI systems
• Recommend practical security controls and guardrails that project teams can implement
• Provide security evidence and technical input for assessment and authorisation activities leading to Authority to Operate
• Contribute to use case assessments, model cards, system cards and assurance cases
• Prepare evidence-based risk advice and reports for project teams, system owners and risk owners
• Brief senior decision makers on technical findings in plain language
• Manage reviews across multiple concurrent projects to fixed delivery timeframes

Required Experience

• Hands-on experience securing AI and machine learning systems, including large language models (LLMs), retrieval augmented generation (RAG) and generative AI applications
• Strong understanding of AI-specific threats and controls, including prompt injection, jailbreaking, data poisoning, model and data leakage, insecure tool or agent integration and guardrail design
• Experience in AI red teaming, adversarial testing or structured evaluation of model behaviour
• Experience working within Defence or Australian Government security environments
• Working knowledge of the Information Security Manual (ISM) and Australian Government security requirements
• Ability to translate technical findings into clear risk advice for technical and non-technical stakeholders, including senior decision makers
• Proven ability to manage multiple concurrent projects and meet fixed delivery timeframes
• Experience supporting Defence or Australian Government assessment and authorisation activities leading to Authority to Operate
• Familiarity with cloud AI platforms such as Azure AI or AWS AI/ML and their security controls
• Knowledge of OWASP Top 10 for LLM Applications, MITRE ATLAS, the NIST AI Risk Management Framework or ISO/IEC 42001
• Experience producing AI governance or assurance artefacts such as use case assessments, model cards, system cards and assurance cases
• Certifications in cyber security, cloud security or AI
• Tertiary qualification in cyber security, computer science, data science, engineering or a related discipline

Security Clearance Requirements

• Must be an Australian Citizen
• Must have a current AGSVA NV1 Security Clearance (minimum)

Working Arrangements

• Onsite
• Hybrid may be considered by exception

Why Profectus

Profectus partners with Australian Government agencies to deliver trusted ICT capability across complex delivery programs. We take a long term, relationship driven approach, supporting our people across the full lifecycle of their engagement while contributing to outcomes of national importance.