Machine Learning Engineer, AI Security Datacenter
Overview
Design experiments that measure the impact of Security Level 5 controls on real ML workflows at a Bay Area AI security nonprofit. Work with AI labs and US intelligence agencies to defend frontier models against nation-state threats. Build the reference tech stack for the first SL5 datacenter, shipping in 2-3 years. Forecast 2028 workloads and run experiments at tractable scales, interpreting results for frontier-scale audiences.
What You'll Do6
- 1Design experiments quantifying how SL5 controls affect real ML research and engineering workflows.
- 2Forecast 2028 frontier workloads covering training, inference, fine-tuning, and interpretability on GPU clusters.
- 3Run experiments at tractable scale, using PyTorch to probe constraints on NVIDIA accelerators.
- 4Interpret experimental results and translate findings into architectural recommendations for frontier labs.
- 5Build data pipelines that collect telemetry from multi-node Kubernetes jobs and S3 storage.
- 6Evaluate storage and orchestration trade-offs for RDMA and NVMe fabrics in high-assurance environments.
Requirements6
- 15+ years training models at an AI lab or equivalent, with current knowledge of frontier methods.
- 2Hands-on experience with PyTorch, multi-node training, and GPU orchestration.
- 3Proficiency with Kubernetes, Docker, and Slurm for managing large-scale jobs.
- 4Familiarity with S3-compatible object storage and high-throughput data pipelines.
- 5Ability to design experiments that yield actionable answers, not just numeric outputs.
- 6Comfortable working at small scale and reasoning about extrapolation to frontier scale.
Salary Insight
Salary not disclosed in listing
Location
Required Skills
Similar open positions
Explore active roles that match your skills and interests.
Security-Level-5
VerifiedNetwork Engineer, AI Security Datacenter
You will own the network fabric and identity layer for the first SL5 datacenter, defending frontier AI models from nation-state threats. The stack includes switch configuration, segmentation, and PKI services, built from the ground up. You will work alongside a lean team of security engineers in San Francisco, with direct input from AI labs and US intelligence agencies. This role ships in 2-3 years, and every decision counts.
Security-Level-5
VerifiedData Center Scaling Lead, SL5 Security
You will own the roadmap for the first SL5 datacenter, scaling from prototype to at-scale build by 2027. You will work with a compact team of engineers and security architects, reporting to the CTO. This role defines the physical and logical constraints that frontier AI labs will build against, with direct input from US intelligence agencies.
Security-Level-5
VerifiedAI Security Fellowship Program Lead
You will own the SL5 fellowship, conferences, and community events, building a global network of senior security experts. You report to the CEO and work with the SL5 ops team to transform the SF office into a security-conscious community hub. This role offers the chance to shape the first SL5 datacenter standards and transition top talent into frontier AI security.
Security-Level-5
VerifiedSenior Director for National Security, AI Datacenter
Own federal and intelligence community partnerships for the first AI datacenter built to Security Level 5, the frontier defense standard for nation-state threats. You'll grow a network of trusted relationships in DC, shape custom whitepapers, and run workshops connecting Bay Area engineers with USG specialists. You'll advise on strategy and accompany the CEO on technical briefings. This role demands deep USG process knowledge and a nonpartisan stance, with an active TS clearance as a bonus. The mission is urgent and the scale is unprecedented: tripling headcount and shipping an at-scale build by 2027.
Unisoft Technology Inc
VerifiedSr Lead AI Data Engineer
Lead design and delivery of AI infrastructure to drive scalable machine learning solutions across enterprise platforms. Own development of end-to-end ML pipelines and foster collaboration between data science and engineering teams. This role differs by focusing on cross-functional leadership and production-grade MLOps implementation.
Security-Level-5
VerifiedStandards Researcher, AI Security & NIST
You will research open questions behind individual Security Level 5 controls, draft control text and overlay content, and maintain the standard as it revises. You will support the Standards Lead in stakeholder engagement with frontier AI labs and US intelligence agencies, writing up decisions from those conversations. This role sits on a small team tripling headcount this year, building the first datacenter purpose-built to Security Level 5. The work has direct impact on defending frontier models from nation-state attackers, and you will shape the reference tech stack that guides labs' buildouts.