Responsibilities
- Lead comprehensive threat modeling across multi-tenant training systems, sandboxed environments, APIs, and internal tools.
- Design and deploy zero-trust networking, identity, and access controls across distributed GPU clusters and cloud environments.
- Create secure-by-default practices for platform teams, covering authentication, secrets management, supply chain security, and container hardening.
- Architect strong tenant isolation and data boundary controls for hosted reinforcement learning workloads where customers execute untrusted code.
- Develop AI-specific security frameworks including model weight protection, training data separation, checkpoint integrity, and gradient privacy.
- Secure the full reinforcement learning training pipeline from sandboxed execution to reward validation and model artifact storage.
- Build defenses against AI-specific threats such as prompt injection, model exfiltration, and adversarial manipulation of training environments.
- Plan and manage external penetration tests across the platform, hosted training systems, and liquid compute infrastructure.
- Establish and run an internal red-teaming program using both automated and manual techniques to test critical systems.
- Oversee vulnerability management including triage, remediation timelines, and root cause investigations.
- Develop security monitoring and alerting systems across infrastructure layers including Kubernetes, distributed clusters, and cloud platforms.
- Implement runtime security controls for containerized training jobs and isolated execution environments.
- Lead incident response efforts, including playbook development, simulation exercises, and post-incident analysis.
- Design and maintain audit logging and forensic capabilities across all customer-accessible systems.
- Drive compliance readiness for SOC 2 Type II and other frameworks required by enterprise clients.
- Own security messaging in customer-facing materials such as security questionnaires, architecture reviews, and trust documentation.
- Collaborate with go-to-market teams to resolve security dependencies and enable enterprise sales.
Work Arrangement
Hybrid — San Francisco
Other
- Flexible work arrangement (remote or San Francisco office)
- Full visa sponsorship and relocation support
- Professional development budget for courses and conferences
- Regular team off-sites and conference attendance
- Opportunity to shape the future of decentralized AI development
Available
