Pinned Loading

1. ForgeML ForgeML Public

   Production-grade MLOps reference covering the full ML lifecycle.

   Python

2. LLM-EDGE LLM-EDGE Public

   Rust LLM Gateway — fast, cost-aware routing across providers with O(1) decision, L1 semantic cache, and adaptive load balancing.

   Rust

3. IvyDoc IvyDoc Public

   Deterministic inference and document extraction for Qwen2.5-3B on Ivy Bridge CPUs. Daemon-based, fixed 1024-token context, PDF→JSON (OCR+RAG), CPU-only.

   Python

4. llama-legacy-gbnf llama-legacy-gbnf Public

   Post-mortem experiment running llama.cpp on legacy hardware, documenting GGUF/GBNF failures (parser, OOM, SIGILL).

   Shell

5. SentinelDrift SentinelDrift Public

   Automated pipeline for monitoring data and concept drift, with feature store, versioning, drift detection, and auto-retraining. Focused on full production lifecycle integrity.

   Python

6. SynapseRT SynapseRT Public

   Stateful recommendation system with incremental online learning for real-time personalization. Designed to optimize latency (P99 < 100ms) and adapt instantly to user feedback without batch retraining.

   Python