01
Network SRE & Automation
Routing, BFD, EVPN/VXLAN, telemetry, fault injection and infrastructure changes delivered through version-controlled workflows.
AN ENGINEERING FIELD LAB / EST. 2026
Production networking.
AI infrastructure.
One engineering mindset.
Documenting the protocols, failure modes and operational details behind modern data center fabrics, high-performance networking and GPU infrastructure.
FABRIC CONTROL PLANE● LAB
SONiC + FRRBGP EVPN / VXLANRoCEv2 / RDMAKUBERNETESDCGM / TELEMETRY
01 / ENGINEERING FOCUS
Built to be tested.
Not just diagrammed.
From control-plane convergence to workload-level impact, the emphasis is reproducible tests and evidence.
02
AI Fabric & RDMA
100GbE networking, RoCEv2, congestion signals and performance measurements, including the limitations surfaced by physical testing.
03
GPU Ops & Kubernetes
GPU Operator workflows, device allocation, NVML-based simulation, DCGM metrics and operational visibility across Kubernetes.
02 / TECHNICAL WRITING
Field notes from
the lab.
Long-form technical articles are being prepared for publication, with architecture diagrams, commands, packet evidence and lessons learned.
FABRIC ARCHITECTUREIN PREPARATION
From Nexus to AI Fabrics: What Actually Changes When Ethernet Meets GPU Networking
AUTOMATIONIN PREPARATION
From Webhook to Production: GitLab CI/CD as the Network Source of Truth
INCIDENT ENGINEERINGIN PREPARATION
RoCEv2 Failure Testing: Congestion, Recovery and the Evidence That Matters
03 / ABOUT THE LAB
Deep network roots.
New infrastructure frontiers.
Pivot AI Lab brings an operations-first perspective from more than two decades in enterprise, telecommunications and data center networking into modern AI infrastructure engineering.
Every published technical claim should survive the same scrutiny as a production incident: documented configuration, measurable behavior and repeatable testing.
Get in touch