Location: Seattle, WA
Cap-Exempt H-1B Position — No lottery required
Persons in these roles are expected to work from our offices in Seattle. On-site requirements vary based on position and team. If you have questions about on-site work arrangements for this role, please ask your recruiter. Our **base** salary range is $176,400 - $264,600, and in addition we have generous bonus plans to provide a competitive compensation package. **Who We Are:** Ai2 is a non-profit research institute at the forefront of open-source AI development. Unlike industry peers, our goal is to share our findings, data, code, and models with the global scientific community. We are seeking a **Director of AI Infrastructure** to oversee the systems that power our research. This leader will be responsible for the full lifecycle of our high-performance computing (HPC) environment which includes on-prem GPU clusters and the software orchestration layer that schedules workloads across a hybrid cloud environment. **Who You Are:** - Systems Expert: You have a deep understanding of the Linux kernel, container runtimes, and distributed systems. You understand the performance implications of InfiniBand topologies and NCCL optimizations. - Strategic Thinker: You look beyond the immediate "fire" to design systems that will scale for the next 3–5 years of AI research. - Pragmatic Leader: You are comfortable making trade-offs between technical elegance and operational necessity. You prioritize reliability and researcher velocity above all else. **Your Next Challenge:** The essential functions include, but are not limited to the following: - Cluster Management: Oversee the availability and performance of dense on-prem GPU clusters. You will partner with hardware vendors and internal teams to ensure our physical infrastructure meets the demands of frontier model training. - Orchestration & Scheduling: Direct the strategy for Beaker, our internal orchestration platform. Your goal is to optimize job scheduling, ensuring high utilization of both on-prem assets and elastic cloud r