Product direction
I set strategic direction for storage and GPU products, working between customer needs, commercial priorities and what engineering can reliably deliver.
Read the professional profileExpertiseLinked to public work
Where I can follow a product question
into the system.
This is not a certification list. Each domain points to public work that shows how I frame the question, test the system and explain what the result changes.
Storage systems
NVMe, NVMe over Fabrics, SPDK, object storage and distributed systems, followed from the workload through the data path and failure model.
See the storage investigationsInfrastructure
Linux, networking, RDMA/RoCE, service lifecycle and the operational details that determine whether a design remains dependable outside a demonstration.
See selected workPerformance testing
I define the workload, configuration and measurement boundary before treating a benchmark number as useful evidence.
Read the field notesGPUs and local AI
Hardware capability, memory architecture, runtimes and model fit. Compatibility is the start of the test, not proof that the software is mature or stable.
Follow the Strix Halo seriesPublic presentation
My Ceph Days London session on hardware choices shows how I connect workload, failure domains and component decisions without pretending there is one correct architecture.
Watch the presentation