AI Security Architect

About Us As a world leading provider of integrated solutions for the alternative investment industry, Alter Domus (meaning “The Other House” in Latin) is proud to be home to 90% of the top 30 asset managers in the private markets, and more than 6,000 professionals across 24 jurisdictions. With a deep understanding of what it takes to succeed in alternatives, we believe in being different - in what we do, in how we work and most importantly in how we enable and develop our people. Invest yourself in the alternative, and join an organization where you progress on merit, where you can speak openly with whoever you are speaking to, and where you will be supported along whichever path you choose to take. Find out more about life at Alter Domus at careers.alterdomus.com The AI Security Architect is responsible for designing, implementing, and governing security frameworks and controls to enable the secure use of AI across the enterprise. This role leads the development of secure-by-design architecture guidelines for AI/ML platforms, defines guardrails for agentic AI workflows and Model Context Protocol (MCP) integrations, and builds observability and detection capabilities to protect AI systems throughout their lifecycle. The ideal candidate combines deep cybersecurity architecture expertise with hands-on experience in AI/ML tooling, LLM security, and enterprise risk management, and is adept at translating complex technical risk into clear stakeholder guidance. Key Responsibilities AI Security Architecture & Design • Develop secure-by-design architecture guidelines for AI/ML platforms, covering data ingestion, model training, deployment, and inference layers. • Define reference security architectures, patterns, and guardrails that enable secure AI development while minimising manual review friction. • Design risk-based, proportionate AI security controls that satisfy regulatory and enterprise requirements while enabling rapid AI development and experimentation. • Partner with engineering teams to automate and embed security guardrails directly into AI development pipelines and toolchains. • Create and maintain guardrails for integrating AI security tools with existing enterprise infrastructure. • Analyse and build security requirements for AI-powered solutions at enterprise scale. MCP Usage, Gateway Integration & Guardrails • Design and enforce security controls for Model Context Protocol (MCP) usage across AI systems and agentic workflows, including tool allowlisting, server trust boundaries, and prompt-level access controls. • Architect MCP gateway integrations with guardrails to prevent unauthorised tool invocation, privilege escalation, and sensitive data leakage across agent pipelines. • Define and implement policies for MCP server authentication, authorisation scopes, and least-privilege access in both internal and third-party MCP deployments. • Implement rate limiting, input/output filtering, content inspection, and sandboxing controls for MCP-enabled agent workflows. • Evaluate third-party MCP plugins and tools for security risks — supply chain threats, data handling practices, and scope creep — prior to enterprise deployment. • Collaborate with platform and AI engineering teams to operationalise MCP guardrail frameworks and ensure consistent enforcement across environments. AI Tools & Ecosystem Integration • Utilise and evaluate AI tools and ecosystems — including OpenAI, Claude, Cursor • Understand AI model architectures, prompt engineering, and the security implications of model configuration and deployment choices. • Work with vendors on AI security tool evaluation, procurement discussions, and implementation. • Integrate AI capabilities into security analytics, dashboards, and workflow automation to enhance detection and response. AI Observability, Logging & Detection • Design end-to-end observability frameworks for AI systems, encompassing prompt/response logging, model behaviour telemetry, agent action trails, and data pipeline audit logs. • Develop detection use cases for AI-specific threats, including prompt injection attempts, jailbreak patterns, sensitive data exfiltration, anomalous agent behaviour, and model drift indicators. • Build security dashboards and KPIs to continuously measure AI security posture and surface emerging risks to leadership. • Establish feedback loops between detection findings and architecture controls to iteratively harden AI systems. AI Threat & Risk Management • Identify and research threats unique to AI systems — model inversion, data poisoning, evasion attacks, data leakage, prompt injection, and adversarial ML attacks. • Lead AI-specific risk assessments and security design reviews across the organization. • Develop adversarial testing frameworks to validate the robustness of AI models and security controls and work with red teams to validate model resilience against adversarial attacks and novel threat scenarios. • Stay current with emerging AI security threats, research, and defensive techniques; present findings to technical and executive stakeholders. Governance, Compliance & Stakeholder Engagement • Establish security policies for ethical and responsible AI use across the organisation. • Ensure compliance with enterprise, industry, and regulatory frameworks including NIST AI RMF, ISO 42001, GDPR, HIPAA, PCI-DSS, and SOC 2. • Collaborate with IT, data engineering, and product teams to embed security into AI development lifecycles. • Lead cross-functional projects to integrate AI security controls across business units. • Develop metrics and KPIs to measure the effectiveness of the AI security programme and report outcomes to executive leadership. Experience Required Qualifications & Skills: • 7+ years in cybersecurity architecture or engineering roles. • 2+ years of hands-on AI/ML development experience using Python and frameworks such as TensorFlow, PyTorch, LangChain, or LlamaIndex. • Proven track record designing and delivering enterprise security frameworks, reference architectures, or architecture patterns at scale. • Experience working in AI/ML or data engineering environments. AI & Machine Learning Security • Strong knowledge of modern AI/ML architectures, pipelines, and tooling — including LLM security, prompt safety testing, and generative AI governance. • Deep understanding of AI-based attacks: model inversion, data poisoning, evasion, prompt injection, and adversarial machine learning. • Expertise in AI red teaming and adversarial robustness testing methodologies. • Knowledge of data science libraries (pandas, scikit-learn, NumPy) and big data technologies. MCP & Agentic AI Security • Understanding of Model Context Protocol (MCP) architecture, including MCP servers, clients, and tool invocation flows. • Experience or familiarity with MCP gateway configuration, guardrail implementation, and access control policy design for agentic AI systems. • Knowledge of security risks specific to multi-agent frameworks and tool-augmented LLM systems. • Understanding of identity scoped permissions and zero-trust for agentic interactions Observability • Experience designing observability and structured logging solutions for AI/ML systems. • Familiarity with cloud-native security monitoring tools Cloud & Infrastructure • Experience with cloud security platforms and AI services across AWS, Azure, and/or GCP. • Strong knowledge of data protection principles and controls required to safeguard training data, model artefacts, and inference outputs. • Understanding of cybersecurity fundamentals: network security, endpoint protection, identity & access management, and threat hunting. Governance & Compliance • Knowledge of regulatory and AI governance frameworks: NIST AI RMF, ISO 42001, GDPR, HIPAA, PCI-DSS, SOC 2. • Familiarity with industry-specific requirements in financial services or healthcare is a strong advantage. Leadership & Communication • Excellent abilit