NVIDIA RTX 4090 / 5090
The workhorses behind local inference and ComfyUI — enough VRAM and throughput to keep multi-model computer-vision pipelines on-prem.
Visit NVIDIA RTX 4090 / 5090 (opens in a new tab)The hardware, tools, and services behind the lab and the projects.
Some links below are affiliate links — I may earn a commission at no extra cost to you. I only list gear and services I actually use.
The workhorses behind local inference and ComfyUI — enough VRAM and throughput to keep multi-model computer-vision pipelines on-prem.
Visit NVIDIA RTX 4090 / 5090 (opens in a new tab)Compact, quiet edge nodes for on-site deployments — small enough to tuck behind a kiosk, capable enough to run real workloads.
Visit ASUS NUC 14 Pro (opens in a new tab)AI-tracking pan/tilt/zoom camera I drive from show-control surfaces for hands-free framing during live events and streams.
Visit OBSBot PTZ camera (opens in a new tab)My go-to show-control surface — Companion turns a Stream Deck into custom button pages that drive Resolume, cameras, and the rest of the live rig.
Visit Elgato Stream Deck + Bitfocus Companion (opens in a new tab)My upgrade from the X1 Carbon — a dual-nozzle, large-format printer for custom enclosures, rack parts, and the hardware behind my edge appliances.
Visit Bambu Lab H2D (opens in a new tab)On-demand GPU cloud I reach for when local cards are saturated — overflow ComfyUI rendering and short training runs without standing up new hardware.
Visit RunPod (opens in a new tab)Payments and checkout built for Singapore businesses — PayNow, cards, and links — that I use to take real money in production.
Visit HitPay (opens in a new tab)