Company Description

Job Description

Scope of Responsibilities

• Research and develop new approaches to AI red teaming, adversarial testing, security evaluation, and robust inference.
• Threat model agentic AI architectures, identifying trust boundaries, attack surfaces, privileged capabilities, and potential failure modes.
• Design adversarial evaluations targeting threats such as prompt injection, indirect prompt injection, tool abuse, privilege escalation, data exfiltration, context or memory poisoning, and unintended agent behavior.
• Build automated security evaluation and regression frameworks that continuously test agents, models, tools, and infrastructure against known and emerging attacks.
• Translate successful attacks and research findings into production mitigations, architectural improvements, and reusable security controls.
• Design secure execution environments for AI agents interacting with tools, code, data, and external systems.
• Build and harden sandboxing and isolation mechanisms for executing agent-generated or otherwise untrusted workloads.
• Design capability boundaries, permission models, and least-privilege access controls for agent tools and services.
• Develop scalable AI infrastructure and services supporting secure model inference and agent execution.
• Own and improve production infrastructure across Kubernetes, AWS, networking, storage, and compute.
• Implement security controls around identity and access management, secrets, network isolation, containers, and service-to-service communication.
• Build scalable APIs, internal platform services, and infrastructure tooling that improve developer productivity, system reliability, and security.
• Improve observability across AI systems through structured logging, metrics, distributed tracing, dashboards, security telemetry, and automated alerting.
• Investigate complex production and security failures across models, agents, distributed systems, and infrastructure.
• Optimize performance, latency, and infrastructure cost while maintaining strong reliability and security guarantees.
• Stay current with emerging attacks against LLMs and agentic systems and rapidly translate relevant research into practical evaluations and defenses.

Qualifications

• U.S. Citizenship is required

Desired Skills: 

• Experience designing or securing AI agent runtimes and tool-execution environments.
• Experience building secure code execution sandboxes, container isolation, microVMs, or other mechanisms for running untrusted workloads.
• Experience with cloud security, infrastructure hardening, IAM, secrets management, network isolation, and zero-trust architectures.
• Experience with offensive security, penetration testing, vulnerability research, or exploit development.
• Experience developing automated adversarial evaluations or integrating security evaluations into CI/CD pipelines.
• Experience with adversarial machine learning, model robustness, or inference-time defenses.
• Familiarity with AI security frameworks and threat taxonomies such as MITRE ATLAS, OWASP guidance for LLM/GenAI applications, or the NIST AI Risk Management Framework.
• Experience securing RAG systems, vector stores, model gateways, or other components of modern AI infrastructure.
• Experience with software supply-chain security and securing model, dependency, and container artifacts.
• Experience working in government, defense, or other high-security environments.