Services · Infrastructure
Observability for GPU infrastructure, built end to end.
C-02 designed and built the observability platform for a GPU cloud provider: scheduler, GPU, and node telemetry, out-of-band management, network fabric, storage, tracing, and logs in one multi-tenant system, plus the diagnostics interface its AI researchers use.
Infrastructure across the team
Where our consultants have built and run infrastructure
Each item is one person's own work, in the roles and at the employers where they did it, cited to their profile.
GPU and HPC clusters
Clusters, schedulers, and supercomputers
- A national supercomputing center's production HPC platform: notebooks, APIs, containers, cluster batch scheduling, and observability one of the twelve · own work
- FFT library optimization with hybrid MPI and OpenMP on a national petascale supercomputer one of the twelve · own work
Observability
Every signal source in one system
Cloud and data platforms
Pipelines and production ML
Security built in
Security in the architecture, not added after
- Finance-grade Go and Rust on multi-region AWS: highest-sensitivity secrets on a no-IP host, egress via a dedicated proxy, systemd sandboxing, and open-source metrics and log-aggregation telemetryC-06
- Five years at the U.S. Department of Energy building auditable, compliance-grade systemsC-02
- Five patents in least-privilege security and permissions risk, two issued and three pending (the individual's)C-01
Scale
Systems at platform scale
Employers named on this site are where our consultants worked, not Transmute Data clients. Patents are the individuals', not Transmute Data's.
How we hand it over
Built inside your operation, run by your team.
We build it inside your operation and hand it over to the team that runs the fleet: runbooks, the monitoring, and its thresholds, and the documentation your engineers need to extend it.
A 30-minute call, on one problem you name.
Shaun takes the call. Name the system. We name the person who has built one like it.