AI and Cloud Security Engineer

Overview

The AI & Cloud Security team is chartered to make that vision operational. We are seeking senior technical owner of the AI security runtime control plane and detection fabric. You will build and integrate the inline enforcement pipeline (AI gateway runtime data protection intent-based enforcement), tune the policies that keep it fast and low-friction, and serve as the final escalation point for the most complex AI and cloud security incidents. You will also help implement and participate in operation of a closed-loop defensive system in which continuous red-team findings automatically improve runtime policies, detection logic, and SOC threat models. You join a team of two AI & Cloud Security Architects and will be the engineering force that turns strategy into operating controls across a multi-cloud estate (Azure, AWS, GCP).

This role involves collaborating with cross-functional teams to ensure the security of AI and cloud environments, AI applications and infrastructure, addressing potential threats, and maintaining compliance with regulatory standards. The ideal candidate will have a strong background in AI technologies and their implementation as well as cloud security, security principles and cybersecurity best practices. The ability to develop automation in both AI and cloud environments with infrastructure as code (IaC) or policy as code principles is required. The engineer will report to AI & Cloud Security leadership and have strong working relationships with IT and application development leadership.

Responsibilities

• Integration build and own the AI runtime control plane. Engineer, deploy and operate the inline enforcement pipeline for AI traffic: an AI gateway chained with runtime data-protection and intent-based enforcement platforms, fronting LLMs, tools, and MCP servers across Azure AI Foundry, AWS Bedrock, GCP Vertex, Copilot Studio, Databricks, and self-hosted environments.
• Harden the gateway layer for resilience : configure buffering and overload management, enforcement timeouts on external-processing filters, layered global/local rate limiting, and prompt-truncation filters to defend against token-exhaustion and “reasoning overload” denial-of-service attacks.
• Integrate the wider AI security ecosystem into one coherent system : onboard AI asset-intelligence context (agent inventory, ownership, blast radius) into enforcement decisions; connect model-security and supply-chain telemetry and continuous red-team signals into SIEM correlation and SOC workflows; stream unified AI detections, decision logs, and policy events for forensics and audit.
• Policy tuning enforce without friction. Author and calibrate runtime policies: prompt-injection and jailbreak prevention, PII detection/redaction and data masking, intent-based filtering, tool-call validation, MCP access control, and agent permission boundaries (including “lethal trifecta” prevention for low-code agents). Balance detection versus hard-block modes to minimize false positives, and tune per AI archetype embedded SaaS copilots, democratized low-code agents, homegrown agentic pipelines, device-based coding agents, and homegrown models.
• Operate the closed loop : translate continuous adversarial-testing findings into runtime policy updates, new detection content, and guardrail tuning, then validate remediation through re-testing so defenses evolve with attackers rather than remaining static.
• L3 coverage final escalation for the AI Security toolset. Lead deep investigations of AI-specific attacks (direct/indirect prompt injection, RAG poisoning, memory poisoning, agent hijacking, MCP/tool abuse, model manipulation, AI-driven data exfiltration), perform decision-replay forensics and root-cause analysis, drive automated containment (agent quarantine, permission revocation, kill switch) toward our sub-15-minute containment objectives, and mentor L2 analysts. Troubleshoot any incidents related to policies.
• Evaluate and pilot AI security vendors : run structured proofs-of-concept for runtime defense, AIDR, AI-SPM, AI red-teaming, and model-security platforms, and provide technical recommendations as the stack matures.
• Engineer and operate cloud security controls across Azure, AWS, and GCP posture policies, detection tuning, and remediation of misconfigurations, excessive entitlements, and workload risks, including the Kubernetes estates (AKS/EKS) hosting the AI gateway layer.
• Design a…