Vaibhav Attre.boot

retro-futurist workstation init

[ 0.000] POST: workstation bus scan complete
[ 0.013] memctl: 64 GiB addressable, ECC nominal
[ 0.029] sched: run queue initialized
[ 0.041] fs: mounting /home/vaibhav
[ 0.068] net: link up on loopback and uplink0
[ 0.091] ui: terminal compositor ready
[ 0.112] session: entering interactive workspace
Loading command bus, filesystem tree, and process monitor.
Vaibhav Attre.ws
/home/vaibhav
shell bus
vaibhav@ws:~$ session
interactive workstation online
type `help` or press `/` to focus the shell
tab completes top match
login shell
/home/vaibhav

Systems and ML infra engineer working on kernels, inference, performance, and end-to-end software.

I build low-level systems when the abstractions need to be understood from first principles, and I build product and infrastructure layers when those systems need to be shipped, measured, and debugged in the real world. Current work ranges from RISC-V operating systems and AI accelerator kernel optimization to edge-cloud inference pipelines and AWS-backed experiment runners.

current base
Redmond, WA
domains
operating systems // ML inference // infra // full-stack systems
recent work
Tenstorrent Blackhole kernels, GoFlyy pipeline, TinyOS, telemetry tooling
operator notes

The through-line in the work is systems discipline: understand the runtime, make state visible, and keep the feedback loop tight enough that performance and correctness are both inspectable.

The portfolio is intentionally organized like a workstation because the projects here are easier to evaluate through interfaces, traces, and technical notes than through generic marketing copy.