Senior Frontier AI Security Engineer

Socket.dev; Mountain View (CA); On-site; USD 207,000 - 300,000; Full time

Benefits offered by this job

Equity

Bonus target

Benefits

Job summary

Google DeepMind seeks a Senior Security Engineer to secure its frontier AI serving infrastructure and platform layer, including multi-tenant data protection, workload isolation, and secure deployment tooling.

The role requires deep background in security operations such as vulnerability triage, incident escalation, and on-call rotations, with focus on preventing risks across Beyond, Agency, and TPU inference systems.

Qualifications

• Bachelor's degree in CS, cybersecurity, ECE or equivalent practical experience.
• 5 years of experience in security engineering/platform security/infrastructure hardening/network protocols/systems security architecture.
• 5 years in security operations (incident response, threat hunting, on-call).
• 5 years of coding experience in system languages (C, C++).

Responsibilities

• Evaluate and harden GDM’s core serving infrastructure (Beyond, Agency, AI Studio APIs) and inference pipelines.
• Develop threat models covering GDM serving layers and inference workloads, including prompt injection mitigation and multi-tenant data protections.
• Build secure-by-default libraries, reference architectures, and deployment templates for secure deployment pipelines.
• Participate in the team's operational security rotation (on-call).

Skills

Security engineering

Security operations

Systems security architecture

Infrastructure hardening

Network protocols

Education

Bachelor's degree in Computer Science or related field

Tools

C

C++

Job description

Google DeepMind seeks a Senior Security Engineer to secure its frontier AI serving infrastructure and platform layer, including multi-tenant data protection, workload isolation, and secure deployment tooling.

The role requires deep background in security operations such as vulnerability triage, incident escalation, and on-call rotations, with focus on preventing risks across Beyond, Agency, and TPU inference systems.