Pinned Loading

1. board-intelligence board-intelligence Public

   Fully offline, GPU-accelerated RAG system for confidential board documents. Hybrid retrieval (BM25 + BGE-M3 dense + RRF + cross-encoder reranking), Qwen3-14B on vLLM, NLI hallucination guard with g…

   Python

2. neuralforge-x neuralforge-x Public

   High-performance, fully-local AI-infrastructure platform — Rust SIMD + rayon vector kernels, a PyO3/maturin Python SDK, a CUDA + Triton GPU engine (Blackwell sm_120), a hand-written HNSW vector da…

   HTML

3. NeuralOps NeuralOps Public

   Production-grade LLMOps platform — Git-style prompt versioning, A/B experiments with LLM-as-Judge scoring, drift detection with Telegram alerts, PII/toxicity guardrails, and a 6-page Streamlit dash…

   Python

4. MemoryMesh MemoryMesh Public

   Production-grade persistent memory server for AI agents. MCP-compatible (Claude Code, Cursor). Four memory types (episodic/semantic/procedural/preference), hybrid vector retrieval (FAISS + bge-larg…

   Python

5. AegisAI AegisAI Public

   Enterprise-grade AI-powered document intelligence platform with autonomous multi-agent workflows, hybrid RAG pipelines, knowledge graph construction, multimodal AI processing, and enterprise-scale …

   TypeScript

6. astraeus-ai-research-engine astraeus-ai-research-engine Public

   Autonomous multi-agent AI research and evaluation engine built with FastAPI, Streamlit, CrewAI and local LLMs (Ollama). Implements structured research generation, verification, critique, scoring, a…

   Python 1