The "Economy Mode" for AI Compute
Self-hosted multi-agent AI platform. Persistent labs, 40 sandboxed tools (code, web, RAG, GPU dispatcher, web3), private Qdrant + LightRAG, sandboxed code execution per lab, anti-loop detection, Ollama/vLLM/OpenAI/Anthropic adapters. Two-command Docker deploy. Apache 2.0.
Distributed document RAG with VRAM-aware workload placement across heterogeneous nodes. Four interfaces: CLI, API, MCP, Web UI.
Legacy version of FlockParser PDF processing system
C# / .NET management service and API for orchestrating local LLM backends, model routing, and GPU resource allocation.
High-performance AI infrastructure marketing site built with Next.js, Tailwind CSS, and Three.js WebGL.
Nimbus is a lightweight, pull-based edge AI workload orchestrator.
Top Distributed Training Platform (Opensource) 🌟 Star if you like it! 🌟
Trust-weighted, cost-optimal Kubernetes scheduler plugin with zero-trust telemetry verification.
Distributed compute orchestration - consensus-aware GPU scheduling across heterogeneous infrastructure
A zero-dependency client load simulator and concurrency stress-testing utility for ZeroGate.
Multi-node PyTorch DDP fine-tuning of a causal LM on Nebius GPU Kubernetes, with SkyPilot workload orchestration — a 2-node torchrun job with verified NCCL collectives.
To associate your repository with the gpu-orchestration topic, visit your repo's landing page and select "manage topics."