I'm a 14-year Cisco DevRel and AI consulting founder (Sierra Code Co) currently going deep on NVIDIA's OpenUSD and Physical AI stack. This site is the journal — friction logs, cross-domain analogies between network automation and 3D scene data, and small demos I'm shipping while I learn.
The format: each entry is a session of learning, written for the next-person-trying-to-do-this. If you came here from a search for "OpenUSD for network engineers" or "Cisco NSO to USD analogy" — that's the point.
2026-06-04
Back from the cruise. Re-running the iPhone-LiDAR-to-Blender pipeline on a hotel room with the slow-down lesson applied. The geometry is dramatically better. The new bottleneck is mesh fragmentation when you try to do anything downstream.
openusdscaniverseblenderphysical-ailearning-log
2026-06-04
Spun up NVIDIA's Isaac Sim on a Brev launchable, trained an Ant RL policy in 2 minutes on an L40S, composed the trained robot with my real-world iPhone hotel scan in a single USD stage. Friction notes throughout.
nvidiaisaac-simisaac-labbrevroboticsopenusdphysical-ailearning-log
2026-06-03
Second of two intermission days. Back to the OpenUSD ramp tomorrow with a pivot toward Isaac Sim and the broader NVIDIA Physical AI stack.
intermissioncisco-livelearning-log
2026-06-02
Two-day gap in the OpenUSD ramp. Cisco Live presentations were scheduled long before this prep started. The honest version of a learning-in-public log includes the days you weren't working on the thing.
intermissioncisco-livelearning-log
2026-06-01
Pivoting from scan-import-render to actually authoring USD scenes in code. Working through LearnOpenUSD's 'Setting the Stage' module, learning what a Stage, Prim, Path, and Type are by typing them out by hand.
openusdpythonlearning-lognetwork-engineer-bridge
2026-05-31
First session of learning OpenUSD: iPhone-scan the cruise promenade with Scaniverse, import the USDZ into Blender, render the first frame. Friction notes throughout.
openusdscaniverseblenderphysical-ailearning-log