Proven expertise in storage performance engineering across file, block, and object protocols (such as parallel filesystems, pNFS/NFS at scale, and S3-scale object stores).
Strong working knowledge of GPU cluster architecture, including RDMA fabrics, GPUDirect Storage, checkpoint/restore dynamics, and resolving bottleneck patterns across compute, network, and storage layers.
Proficiency with industry benchmark standards (MLPerf Storage, IO500) and load-generation tools (elbencho, fio, vdbench), alongside the capability to design custom, workload-representative tests.
Demonstrated ability to perform architecture-level competitive analysis, along with a firm grasp of multi-tenant resource management concepts (QoS, rate limiting, and workload isolation) in distributed systems.