About this role
This is an architect-level, fully remote cybersecurity role focused on AI security for a healthcare insurance client. You will design enterprise AI security architecture for locally hosted LLM models, define guardrails, and implement governance and security standards across the AI lifecycle.
Key Responsibilities
- Design and implement enterprise AI security architecture for locally hosted AI/LLM platforms
- Define and enforce AI guardrails including prompt injection prevention and output validation
- Perform AI threat modeling for LLM-specific vulnerabilities such as poisoning and adversarial inputs
- Integrate security controls across AI lifecycle including data ingestion, model training, inference, and output handling
- Establish security metrics, architecture governance, and maintain architecture documentation
Technical Overview
The role emphasizes hands-on AI security design: threat modeling for LLM vulnerabilities, enforcing prompt injection prevention, and applying controls across data ingestion, model training, inference, and output handling. You will create security blueprints, standards, design patterns, and metrics to monitor the effectiveness of the enterprise AI security posture.
Ideal Candidate
The ideal candidate is a senior cybersecurity architect with 10+ years of overall security experience and at least 2 years of hands-on AI security architecture experience. They have designed enterprise AI security for locally hosted (non-SaaS) LLM systems, including guardrails, prompt injection prevention, and end-to-end threat modeling and control implementation across the AI lifecycle.
Must-Have Skills
10+ years of overall cybersecurity experience with strong focus on architecture and design2+ years of currenthands-on experience in AI security architecture (MANDATORY)Proven experience securing locally hosted AI/LLM models (non-SaaS environments)Define and enforce AI guardrailsincluding prompt injection preventioninput/output filteringresponse validation and policy enforcementPerform AI threat modeling and identify vulnerabilities specific to LLMsincluding model misusepoisoningand adversarial inputsIntegrate security controls across AI lifecycleincluding data ingestionmodel traininginferenceand output handling
Required Skills
AI security architecturelocally hosted AI/LLM modelsguardrailsprompt filteringoutput validationprompt injection preventioninput/output filteringresponse validationpolicy enforcementAI threat modelingmodel misusepoisoningadversarial inputssecurity controls across AI lifecycledata ingestionmodel traininginferenceoutput handlingsecurity blueprintsarchitecture standardsdesign patternsarchitecture governancesecurity metricsbuild vs. buy
Hard Skills
AI security architecturelocally hosted AI/LLM modelsguardrailsprompt filteringoutput validationprompt injection preventioninput/output filteringresponse validationpolicy enforcementthreat modelingAI threat modelingsecurity controls across AI lifecycledata ingestionmodel traininginferenceoutput handlingsecurity metricsenterprise architecture standardsarchitecture governancesecurity blueprintsdesign patterns for AI and machine learning platformsarchitecture documentation
Soft Skills
Technical leadershipCollaboration with engineeringdata scienceand infrastructure teamsArchitecture governance participationGuidance and mentoringTechnical documentation and operational procedure ownership
Keywords for Your Resume
Cybersecurity ArchitectAI SecurityLLMLarge Language ModelsGuardrailsenterprise AI security architectureprompt injection preventionprompt filteringoutput validationinput/output filteringresponse validationpolicy enforcementAI threat modelingmodel misusepoisoningadversarial inputsAI lifecycledata ingestionmodel traininginferenceoutput handlingsecurity blueprintsarchitecture standardsdesign patternsarchitecture governancesecurity metricsenterprise security posturesecure AI solution designbuild vs. buyAI security architectureguardrails
Deal Breakers
Must have 2+ years of current, hands-on experience in AI security architecture (MANDATORY), Must have proven experience securing locally hosted AI/LLM models (non-SaaS environments), Must be able to meet the work authorization requirement: permanent residency or citizenship
Get matched to jobs like this
Luna finds roles that fit your skills and career goals — no endless scrolling required.
Create a Free Profile