Services
Creative projects and tools — each with its own frontend, pipeline, and agent
nerf-miner
3D reconstruction pipeline — photo scanning, CLIP embeddings, gaussian splatting
active
Phase 2 DONE — first splatting job completed on 5090. Phase 3: service integration
Next
›Finish last 2 photo embeddings (GoPro videos on NAS)
›Begin candidate review (224 pending)
›Fix TrainingResult Pydantic model (metrics Dict[str, float] → Dict[str, Any])
+3 more
5 completed
nova
AI assistant API — conversational AI, knowledge base, chat interface
active
Phase 2 complete. Phase 3: modernise.
Next
›Test end-to-end chat flow
›Assess frontend state (nova-client, easia)
›Phase 3 — modernise (replace OpenAI with local LLM, remove Stripe/AWS)
2 completed
Resources
Compute machines, tunnels, and SSH routes
Work Machines
1/5
tunnels up
Home Devices
3
on LAN
Total GPUs
6
across all machines
Services
3
tracked
RTX 5090
work-5090
online
NVIDIA RTX 5090 · 32GB
gpucomputewindowswork
SSH
ssh -i ~/.ssh/workcloud -o IdentitiesOnly=yes -p 10001 phoria@localhost
Shell: bash (WSL)
RTX 5080
work-5080
offline
NVIDIA RTX 5080 · 16GB
gpucomputewindowswork
SSH
ssh -i ~/.ssh/workcloud -o IdentitiesOnly=yes -p 10002 sam@localhost
Shell: powershell
Ada Render 1
work-ada-1
offline
NVIDIA Ada
gpurenderwindowswork
SSH
ssh -i ~/.ssh/workcloud -o IdentitiesOnly=yes -p 10003 sam@localhost
Shell: powershell
Last updated: 10:44:01 PMhme.samtg.xyz — home services →