Go and Rust internals, Istio and Envoy service mesh architecture, GPU sharing mechanics (cGPU, HAMi, MIG), and LLM gateway design patterns.
Currently hacking on
Kompakt, an open-source Kubernetes bin-packing admission controller built on scheduling gates. GPU inventory and LLM gateways.
Available for
Discussing AI infrastructure, Kubernetes cost optimization, GPU scheduling, and LLM platforms. Open to OSS collaboration on Kompakt and technical writing opportunities.