Senior AI SecOps Engineer

1. Role Overview

We are seeking a Senior AI SecOps Engineer to design and establish a practical security framework for enterprise AI agents and LLM-powered applications.

The role focuses on identifying AI-specific security risks, defining and implementing security controls, establishing agent governance, and validating that controls operate as intended. The engineer will work closely with cloud security, application engineering, platform, and governance teams to translate security requirements into implementable controls and operational practices.

This is a hands-on consulting role requiring a balance of security architecture, engineering implementation, risk assessment, and technical documentation. The successful candidate will establish the initial framework, develop reference implementations for selected controls, validate them against representative AI use cases, and enable the customer's engineering and security teams to maintain and extend the solution.

Engagement details

• Duration: 6 months
• Time commitment: Approximately 60100 hours per month
• Engagement type: Part-time consulting / contract
• Primary focus: AI agent security, security controls, governance, and risk management
• Working model: Collaboration with the customer's existing engineering, cloud security, and governance teams

2. Key Responsibilities

A. AI Agent Security Architecture

• Assess the security posture of existing or planned AI agents and LLM-powered applications.
• Develop threat models covering prompt injection, unauthorized tool execution, excessive agent permissions, sensitive-data exposure, insecure integrations, and untrusted inputs or outputs.
• Define security boundaries and trust relationships between agents, users, tools, APIs, enterprise data sources, and external services.
• Establish security patterns for agent identity, authentication, authorization, secrets management, least-privilege access, and controlled tool execution.
• Define appropriate approval and escalation requirements for high-impact or sensitive agent actions.
• Align AI-specific security controls with the customer's existing cloud and application security architecture.

B. Security Controls and Reference Implementations

• Define a reusable baseline of security controls for enterprise AI agents, categorized by risk, autonomy, data access, and business impact.
• Establish implementation patterns for access control, tool permissions, input/output validation, sensitive-data handling, and policy enforcement.
• Develop or configure reference implementations for selected controls within the customer's existing technology stack.
• Define secure integration patterns for agent frameworks, APIs, enterprise systems, and relevant agent communication protocols.
• Evaluate controls against realistic attack scenarios and document limitations, residual risks, and remediation recommendations.
• Collaborate with engineering teams to incorporate security requirements into the AI application development lifecycle.

C. AI Governance and Risk Management

• Define an inventory model for AI agents, capturing ownership, purpose, business criticality, data access, connected tools, permissions, and autonomy level.
• Establish a risk-classification approach to determine appropriate security requirements for different agent types.
• Define minimum security requirements for agent onboarding, deployment, modification, and retirement.
• Map relevant AI security controls to established security and risk-management frameworks, such as the NIST AI Risk Management Framework, NIST Cybersecurity Framework, and OWASP guidance for LLM applications, where applicable.
• Establish processes for recording exceptions, assigning control ownership, tracking remediation, and maintaining audit evidence.
• Support security and governance stakeholders in defining accountability and review processes for high-risk agents.

D. Security Validation and Testing

• Develop a practical test plan to validate security controls against representative AI agent threat scenarios.
• Conduct or coordinate security testing covering prompt injection, unauthorized tool calls, privilege misuse, sensitive-data leakage, and policy bypass attempts.
• Validate authorization boundaries, tool restrictions, data-access controls, and approval mechanisms.
• Document test results, control effectiveness, identified gaps, and remediation actions.
• Recommend improvements based on test outcomes and residual risk.
• Establish repeatable security validation procedures that the customer's engineering team can use for future AI agent deployments.

E. Operational Readiness and Handover

• Define security review procedures for new and modified AI agents.
• Develop operational runbooks for security assessments, control exceptions, remediation tracking, and escalation.
• Establish a practical process for reviewing agent permissions, security configurations, and compliance with the approved security baseline.
• Transfer knowledge, implementation artefacts, and operational responsibilities to the customer's designated teams.
• Provide recommendations for future improvements to the AI agent security framework.

3. Expected Deliverables

Deliverables will be prioritized and agreed with the customer during the engagement kickoff.

Deliverable

Expected outcome

AI Security Assessment

Documented architecture review, threat model, risk register, and prioritized recommendations

AI Agent Inventory and Risk Model

Reusable inventory schema, ownership model, risk classification, and security requirements

AI Agent Security Baseline

Documented minimum controls and implementation guidance for different risk levels

Security Reference Implementations

Working examples of selected controls integrated with representative agents or applications

Security Validation Pack

Test scenarios, execution evidence, findings, and remediation recommendations

Governance and Evidence Framework

Control mapping, exception tracking, evidence requirements, and review procedures

Operational Handover Pack

Runbooks, architecture documentation, implementation artefacts, and knowledge-transfer materials

Executive Summary

Final assessment of security posture, outstanding risks, completed deliverables, and recommended next steps

The scope of reference implementations and production integrations will depend on the existing environment, platform readiness, access, and available engineering support.

4. Required Qualifications

Experience

• 8+ years of overall experience in cybersecurity, cloud security, application security, or security engineering.
• 5+ years of relevant hands-on experience implementing or assessing security controls in enterprise applications, cloud environments, or distributed systems.
• Demonstrable experience with security architecture, threat modelling, identity and access management, and secure system integrations.
• Hands-on experience securing at least one production LLM application, AI agent, or comparable AI-enabled system. Substantial implementation experience in AI security research or security engineering may also be considered.
• Experience translating security requirements into working controls, implementation guidance, and testable acceptance criteria.
• Experience collaborating with application engineers, platform teams, security stakeholders, and technical leadership.

Technical Skills

AI and application security

• Understanding of LLM and agent architectures, including model interactions, tool calling, retrieval-augmented generation (RAG), and external system integrations.
• Understanding of AI-specific threats, including prompt injection, insecure tool execution, sensitive-data disclosure, excessive agency, and trust-boundary failures.
• Ability to assess agent permissions, authorization boundaries, data access, and the potential impact of autonomous actions.
• Knowledge of AI security testing, guardrails, and the limitations of prompt-based safeguards.

Cloud, identity, and security engineering

• Strong understanding of IAM, least-privilege access, workload identity, secrets management, API security, and secure integration patterns.
• Experience with at least one major cloud platform, such as AWS, Azure, or Google Cloud.
• Understanding of encryption, access control, audit requirements, vulnerability management, and secure software development practices.
• Ability to design and validate security controls that complement existing enterprise security tools and processes.

Governance and risk management

• Ability to translate security risks into control requirements, risk classifications, evidence requirements, and remediation plans.
• Familiarity with AI governance and security frameworks, such as NIST AI RMF, OWASP Top 10 for LLM Applications, and relevant organizational security standards.
• Ability to document risk assessments, control mappings, exceptions, testing evidence, and remediation plans.

Integration and implementation

• Working knowledge of REST APIs, JSON, and common integration patterns.
• Ability to review implementation code, troubleshoot integrations, and work with engineering teams to validate security controls.
• Familiarity with Git, CI/CD, and security testing in the software development lifecycle.

5. Desirable Qualifications

• Experience with agent frameworks, agent orchestration platforms, MCP (Model Context Protocol), or similar tool-integration mechanisms.
• Experience with AI security evaluation tools, red-teaming approaches, policy enforcement, or agent security testing solutions.
• Experience with cloud-native AI services and enterprise LLM platforms.
• Experience in regulated or compliance-sensitive environments.
• Familiarity with Python, SQL, or scripting for security testing and automation.
• Experience integrating security controls into CI/CD pipelines.

Python, SQL, and specific data-platform technologies are not mandatory requirements unless the agreed implementation scope explicitly requires custom development.

6. Candidate Expectations

The successful candidate is expected to:

• Operate as a hands-on security engineer rather than a documentation-only consultant.
• Make pragmatic recommendations that can be implemented within the customer's existing architecture and constraints.
• Distinguish immediate security priorities from longer-term improvements.
• Demonstrate how proposed controls work through reference implementations, configuration, or repeatable tests.
• Communicate technical risks, trade-offs, residual risks, and dependencies clearly.
• Work independently while coordinating with the customer's engineering and security teams.
• Produce reusable documentation and implementation artefacts that the customer can maintain after the engagement.
• Deliver against an agreed backlog, milestones, and acceptance criteria within the available consulting hours.

Ideal candidate profile: A senior cloud or application security engineer with demonstrable AI/LLM security implementation experience, strong security architecture fundamentals, and the ability to turn security requirements into working controls and repeatable validation tests.