Ai Security Architect

Do you know
askblue?
We were born in ****, and we provide services in the field of information technology.
We are looking for
AI Security Architect
to join our company, in one of our projects in
Lisbon
,
in hybrd regime
.
Tasks:
Support and review AI/GenAI use cases from design through production;
Define security requirements for AI platforms, APIs, agents, tool use, and integrations;
Perform architecture reviews, threat modelling, and security reviews of AI solutions;
Validate integrations and deployments against security, privacy, logging, and auditability requirements;
Identify and mitigate risks such as:
Prompt injection / jailbreaks;
Sensitive information disclosure / data leakage;
Unsafe tool use / excessive agency;
Abusive or unbounded consumption;
Insecure integrations or poisoned inputs;
Define guardrails for data usage, especially for sensitive, personal, confidential, or regulated data;
Ensure secure logging, monitoring, audit trails, and evidence of control effectiveness;
Contribute to AI security governance, standards, and secure-by-design patterns.
Requirements:
Background in cybersecurity, ideally application security, product security, cloud security, or security architecture;
Understanding of GenAI / LLM-based solutions from a security and risk perspective;
Experience with security reviews, architecture reviews, threat modelling, and defining technical controls;
Experience with IAM, API security, secrets management, logging, monitoring, and control validation;
Knowledge of data protection, privacy, data minimization, and secure handling of sensitive or regulated data;
Ability to challenge technical implementations and validate that controls are effectively applied;
(Preferred) Familiarity with AWS and/or Azure environments in the context of AI workload security;
(Preferred) Exposure to AWS Bedrock, Azure OpenAI / Azure AI Foundry, or similar GenAI platforms, especially around:
Guardrails and content controls;
IAM and least privilege;
Logging, observability, and auditability;
Data protection and sensitive data handling;
Familiarity with Databricks as a platform to review from a security perspective, including: Access control and data permissions, Workspaces, jobs, pipelines, and notebooks, Secrets, networking, and data governance.
Familiarity with:
OWASP Top 10 for LLM Applications / Agentic AI;
SAIF (Secure AI Framework);
Model Context Protocol (MCP) and agent-to-tool security considerations;
Agent-to-Agent (A2A) architectures and agentic trust boundaries;
Emerging risk taxonomies such as MCP-38;
Exposure to DevSecOps and/or MLOps as a collaboration context.
Core Skills:
Can translate AI security risks into clear technical requirements;
Strong risk identification, prioritisation, and decision-making;
Able to challenge and validate implementations effectively;
Works well with security, product, engineering, and data teams;
Focused on practical, enforceable, and auditable controls.
Important Note:
This is a cybersecurity role applied to AI/GenAI. It is not primarily a model development, data engineering, or operational ownership role for pipelines, notebooks, or ML platforms.
Work Arrangement:
Hybrid (On-Site 2x per week)
Offer:
Health Insurance;
3 and a half days of leave per year + 22 vacation days;
Unlimited access to Udemy.
If you are interested in the opportunity, upload your C.V.
askblue – where business meets technology