Lead AI Security Engineer | Agentic SOC

NOTE: This role does not support Sponsorship The Opportunity We’re partnering with a leading enterprise software company to hire a Lead AI Security Engineer to help architect and build an Agentic Security Operations Center from the ground up. This is not a traditional cybersecurity, SOC analyst or DevSecOps role. It is a deeply technical AI/software engineering position embedded within security , designed for someone who has already built production applications powered by Large Language Models and wants to apply that expertise to complex cybersecurity problems. The goal is ambitious: move beyond reactive alerting and traditional security automation toward autonomous, context-aware AI systems capable of triage, semantic correlation, investigation and proactive threat mitigation. You’ll work alongside security engineers, Tier 3 analysts and threat hunters to build the intelligent infrastructure behind that vision. What You’ll Build This is a hands-on engineering role. You’ll architect, code, test and deploy production systems, including: Autonomous AI Agents: Design and deploy autonomous and semi-autonomous agents that interact directly with enterprise SIEM, EDR/XDR and threat-intelligence platforms. Multi-Agent Systems: Architect sophisticated agentic workflows using DAGs, routing, tool calling and multi-agent orchestration. SOC Automation: Translate human-driven security playbooks and analyst workflows into deterministic and heuristic AI pipelines. Production RAG: Build and optimize RAG systems that incorporate security telemetry, network topology, threat intelligence and historical incident data. LLM Performance Engineering: Evaluate and benchmark open-source and commercial models across accuracy, context utilization, latency, scalability and cost. Human-in-the-Loop Systems: Engineer safe escalation paths between autonomous agents, Tier 3 analysts and threat hunters. Secure AI Architecture: Build safeguards against prompt injection, data poisoning, tool abuse and other emerging LLM vulnerabilities. Cloud-Native AI Infrastructure: Deploy and scale AI workloads within containerized cloud environments. What You Must Bring We want to be extremely clear about the technical bar for this position. Required: 4+ years of professional software engineering experience 1.5+ years specifically building applications powered by Large Language Models 5+ years programming professionally with Python 2+ years deploying scalable, responsible AI solutions on AWS, Azure or GCP Demonstrated experience building production-grade AI systems , not simply prototypes, prompt engineering or API integrations Strong understanding of software architecture, asynchronous systems and production engineering Ability to personally architect and code sophisticated AI solutions Ability to turn ambiguous security problems into concrete technical architectures and working systems Ability to work on-site five days per week in McLean/Tysons, Virginia Experience That Will Set You Apart We’re especially interested in engineers with hands-on experience in several of the following: LangChain, LlamaIndex, AutoGen, CrewAI or Semantic Kernel Multi-agent architectures and orchestration Production Retrieval-Augmented Generation systems Pinecone, Qdrant, Milvus, Weaviate or comparable vector databases Splunk, CrowdStrike, Microsoft Sentinel, Palo Alto XSOAR or similar security platforms REST APIs and webhooks within cybersecurity environments Kubernetes, EKS or AKS LLM evaluation and performance optimization Fine-tuning methodologies including LoRA and QLoRA Open-source models such as Llama or Mistral Security operations, threat intelligence or incident response AI security, guardrails and secure agent/tool execution Who Will Thrive Here You don't just follow developments in AI. You have already put AI systems into production, watched them encounter the messy realities of production environments, and learned how to make them more reliable, scalable and secure. You are deeply technical and still want to write code and build systems . You’re comfortable operating where the architecture isn't already defined. You can take a problem such as “How could an autonomous agent safely investigate and respond to this threat?” and turn it into an architecture, determine the appropriate models and tools, build it, test its boundaries and ultimately ship it. You also understand that autonomous AI inside a security environment requires an exceptionally high bar for reliability and control. What an agent is allowed to do matters just as much as what it is capable of doing. If you want a role where the blueprint already exists, this probably isn't it. If you want to help write the blueprint for how AI changes enterprise security operations , we'd like to meet you. Compensation Actual compensation will depend on experience, technical depth and qualifications. The total rewards package also includes a discretionary performance bonus, comprehensive benefits and potential eligibility for equity awards. Location This position is on-site four to five days per week. Candidates must be able to work from the McLean-area office on a full-time basis. #J-18808-Ljbffr AlignityX