Sr AI Engineer, Security

Build and lead AI security for agent and model platforms across AWS and GCP, focused on practical controls, guardrails, and incident readiness.

Responsibilities

• Secure the AI platform across AWS and GCP.
• Design, enforce, and operationalize guardrails for AI workloads on:
  • AWS using Bedrock, Lambda, IAM, KMS, SCPs, GuardDuty, and Security Hub.
  • GCP using Vertex AI, IAM, VPC Service Controls, Org Policies, Security Command Center, and Cloud KMS.
• Define how agents and non-human accounts are identified and authorized using:
  • Short-lived, narrowly scoped credentials
  • Delegated user permissions
  • A per-agent audit trail to limit blast radius for misbehavior
• Build agent containment including separate read and write permissions, network egress controls, sandboxed execution, approval gates for high-risk actions, and kill switches.
• Design an agent gateway or control plane to keep agents and tool calls within approved limits.
• Secure the AI supply chain by governing model and dataset intake, including scanning for unsafe formats such as pickle; vetting model and dataset sources from public hubs such as Hugging Face; approving new MCP servers and agent tools; and maintaining an AI-BOM inventory of models, data, and tools used by each system.
• Defend against prompt injection, tool poisoning, RAG poisoning, and sensitive data leakage.
• Implement detection and logging for agent and model activity; own AI-specific incident response playbooks.
• Lead threat modeling for new products, features, and internal solutions spanning consumer experiences to internal tools and agent workflows, using frameworks such as STRIDE, OWASP LLM/GenAI Top 10, MITRE ATLAS, and OWASP ASVS; translate outcomes into design requirements and release criteria.
• Review pull requests, architecture designs, and release readiness for apps to agent fleets, providing concrete suggestions that move work forward.
• Create paved-road frameworks, reference architectures, and secure-by-default templates that prevent classes of risk rather than handling issues one PR at a time.
• Embed security into CI/CD with secrets scanning, SAST/SCA, GitHub Actions hardening, and policy-as-code in Terraform.
• Run an ongoing red-team and evaluation program for models, agents, and LLM apps; test AI features before launch.
• Partner with the global cybersecurity team to align Realtor.com AI and cloud security with group standards, policies, and risk frameworks.
• Coordinate across News Corp businesses on incident response, threat intelligence, and vulnerability disclosure.
• Share AI security patterns, threat models, and lessons learned with other News Corp businesses and bring insights back to Realtor.com.
• Build agents to support security work, including triage, alert investigation, IAM least-privilege audits, and creating tested, secure-by-default fix PRs.
• Use Claude Code, Devin, and MCP to accelerate the security team and share reusable patterns with other engineering teams.
• Track emerging AI threats, incidents, research, and vendor capabilities, converting them into platform actions.
• In major industry incidents, assess exposure quickly and lead response coordination.
• Prototype new controls when best practices do not yet fit, run small experiments, measure results, and roll improvements into platform standards.
• Write internal briefings on major developments and recommend when to adopt, pilot, or skip tools and practices.
• Create and run AI security training for engineers, product teams, and the wider company covering prompt injection, data leakage, excessive agency, and secure use of tools like Claude, Glean, Devin, and Gemini.
• Act as the AI security voice in platform strategy: evaluate vendors and models, set expectations for production readiness, and shape AI governance across the AI lifecycle.
• Brief leadership on AI risk and where to invest.
• Represent Realtor.com externally via conference talks, open-source work, or industry working groups.

Requirements

• 5+ years in cybersecurity, cloud security, AppSec/product security, or security platform engineering.
• Hands-on depth in AWS security plus real experience in GCP (or strong multi-cloud experience).
• Security thinking grounded in permissions, scopes, logs, approval gates, data boundaries, and rollback rather than relying on “do not do X” prompts.
• Experience securing LLM or agentic systems, including prompt injection defenses, retrieval and data-boundary risks, agent identity, and model or tool supply-chain integrity.
• Threat modeling experience for products and platforms (not only infrastructure), including leading sessions with engineers and product managers and turning output into requirements they follow.
• Hands-on experience building with agents, tool calling, or MCP, and with AI coding tools.
• Strong code-review skills in at least one of Python, TypeScript, or Go, and comfort with Terraform/IaC.
• Experience with zero trust, least privilege, SSO/OIDC/OAuth, secrets management, and vulnerability management.
• Experience working within a larger security organization (parent company, central security team, or federated model) with the ability to balance local execution speed with enterprise alignment.
• Stay current on AI security research, incident write-ups, and community learning; apply something recent.
• Builder mindset: prototype new controls and operate where playbooks are still forming.
• Ability to explain risk to any audience, from PR comments to executive briefings to company-wide training.

Technologies

• AWS, GCP
• Bedrock, Lambda, IAM, KMS, SCPs, GuardDuty, Security Hub
• Vertex AI, VPC Service Controls, Org Policies, Security Command Center, Cloud KMS
• Hugging Face, MCP
• Claude Code, Devin, Claude, Glean, Gemini
• Python, TypeScript, Go
• Terraform
• CI/CD, GitHub Actions, SAST, SCA
• STRIDE, OWASP LLM/GenAI Top 10, MITRE ATLAS, OWASP ASVS, RAG, pickles

Benefits

• Inclusive and competitive medical, Rx, dental, and vision coverage
• Family forming benefits
• 13 paid holidays
• Flexible time off
• 8 hours of paid volunteer time off
• Immediate eligibility into Company 401(k) plan with 3.5% company match
• Tuition reimbursement program for degreed and non-degreed programs
• 1:1 personalized financial planning sessions
• Student debt retirement savings match program
• Free snacks and refreshments in each office location

Nice to Have

• AWS Security Specialty, Google Professional Cloud Security Engineer, CCSP, CISSP, OSCP, or GIAC (GCSA/GWEB)
• Experience red-teaming or evaluating AI models or agents
• Familiarity with NIST AI RMF, ISO 42001, Google SAIF, CoSAI, NIST SSDF, or the EU AI Act
• Experience securing Snowflake (RBAC, masking, network policy)
• Conference talks, published research, or open-source security contributions

Work Location and Work Style

• Onsite in Austin, TX, United States
• In-person collaboration is emphasized; for most roles, employees work four or more days in the office