PRAXIS Labs — Enterprise Production Testing Interest
Hey,
My name is Michael Clarke, founder of Noetic Synthesis LLC and PRAXIS Labs — Research Development, Cherry Hill, New Jersey. GS1 registered commercial operation, GLN: 1200202171808.
We are running your Pantheon-RP-1.8-24b-Small-3.1 in production on an enterprise Dell PowerEdge cluster — multiple nodes over a Dell S6010-ON 40GbE RoCEv2 fabric, just under 200GB total GPU VRAM. Full pipeline: llama.cpp inference, BGE-large embeddings, Milvus, Neo4j, PostgreSQL memory stack, Unreal Engine 5 MetaHuman visual presence, and live SMS and WhatsApp delivery via ADB. Not a demo environment.
This week we deployed a Qwen3-35B-A3B consultation layer that silently queries Qwen before Pantheon generates — complexity routing plus keyword triggers. Working in production as of today.
We found Pantheon-Proto-RP-1.8-30B-A3B. The Qwen3 MoE base is the direction we want. PRAXIS Labs can test it at a scale and in a pipeline architecture most of your community testers don't have access to. Our research is closed and private.
If you or your team want serious production feedback from enterprise hardware, we are interested.
Michael Clarke
Founder, Noetic Synthesis LLC / Praxis Collective
PRAXIS Labs — Research Development
7 Katherine Avenue, Cherry Hill, NJ 08002
architect@noeticsynthesis.ai
856-482-5419
GS1 Registered — GLN: 1200202171808
GitHub: https://github.com/Praxis-Collective/Praxis (closed / private research)
