Principal AI Security Engineer
Cerebras builds wafer-scale AI computing systems and a cloud inference platform for training, fine-tuning, and serving AI models.
About Cerebras Systems, Inc.
Cerebras Systems is an AI-infrastructure company founded in 2015. It sells rack-scale wafer-scale computing systems and provides cloud-based, API-accessible AI inference alongside on-premises deployments.
Skills
About the Role
You will lead hands-on security engineering across AI platforms, enterprise IT, infrastructure, and agentic systems. You will design practical security architecture and reusable controls for training, inference, model serving, customer workloads, identity, runtime environments, data, and developer platforms. You will automate policy validation, reviews, evidence collection, remediation, and telemetry-driven response workflows.
Requirements
- 10+ years of experience in security engineering, platform security, infrastructure security, product security, or related technical security roles
- Hands-on engineering ability in Python and at least one additional production language
- Experience designing, building, operating, and improving security controls as code
- Cloud and infrastructure security experience, preferably with AWS
- Understanding of IAM, networking, secrets management, logging, and cloud-native control planes
- Understanding of SSO, MFA, OAuth, service accounts, workload identity, authorization, privileged access, and least privilege
- Experience securing containers, Kubernetes, isolated workloads, secure development environments, distributed compute platforms, or production service infrastructure
- Familiarity with AI security, LLM application security, agentic workflows, MCPs, prompt injection, autonomous coding agents, or AI platform security
- Written communication skills and ability to influence senior technical stakeholders
Responsibilities
- Define security architecture and build controls for AI platforms, training and inference workflows, model-serving systems, customer workloads, developer workflows, and agentic systems
- Develop reusable AI and agent security patterns for identity, authorization, delegated authority, tool access, MCPs, connectors, secrets, approvals, isolation, and auditability
- Design runtime controls that constrain execution, access, data exposure, model and tool interaction, and blast radius
- Build security capabilities as code using infrastructure as code, configuration as code, policy as code, GitOps, CI/CD, and automated validation
- Define secure development patterns for AI systems, agents, prompts, tools, models, policies, evaluations, releases, and rollback
- Automate security reviews, policy checks, evidence collection, control validation, and remediation
- Instrument AI, agent, and platform activity with telemetry, traceability, policy decisions, audit logs, anomaly signals, and response workflows
- Lead security reviews and influence architecture through practical design changes and reusable controls