UATC-Closed-loop VRAM control and dynamic data pruning for LLM training
Similar Projects
TruLayer – tracing, evals, and a control loop for production LLMs
Automated rollback on regression is a killer feature LangSmith doesn't have.
SCP – A protocol that drops LLM API calls to zero in 60fps physics loop
Subsumption Architecture revival cuts LLM calls with pattern cache misses.
Looplet – a 0-dep agent loop you own
Iterator-first design beats black-box frameworks like LangChain for debugging.
BNNR – a closed-loop pipeline for improving vision models
XAI-driven model improvement loop, but Weights & Biases already tracks experiments better.
I just shipped the canonical neuro-symbolic control demo
Compiles stochastic Petri nets into a verified SNN (LIF + bitstream) and executes closed-loop replay on real DIII‑D shot data with liveness/boundedness proofs and SHA256 proof bundles. The notebook also ships side-by-side SNN vs PID vs MPC metrics, a FusionKernel digital twin, and deterministic artifact export — impressive technical depth for experimental fusion control, but clearly aimed at specialists and not turnkey for newcomers.