Skip to content
View Turbo31150's full-sized avatar
🧠
Building Multi-GPU AI Clusters
🧠
Building Multi-GPU AI Clusters

Block or report Turbo31150

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Turbo31150/README.md

License: MIT

🔮 ALKYMIA — Startup IA & Systèmes Distribués Souverains

Fondateur : Franck Delmas | Toulouse, France

Site LinkedIn


⚡ JARVIS OS — Infrastructure IA On-Premise

1 000+ agents autonomes · 6 machines · 12 GPUs · P95 18ms · 0€ cloud · Uptime 100%

Métrique Valeur
Agents simultanés 1 000+
Latence P95 18ms
Coût cloud 0€/mois
Lignes Python prod 280 741
Chaînes Domino 835
Commandes vocales Whisper 2 658 (<300ms)

🚀 Applications ALKYMIA

App Description
Phanesis Moteur de matching sémantique & graphe de connaissances
Pneuma Assistant IA vocal on-premise (STT/TTS souverain)
Quintessence Distillation & optimisation de modèles LLM
Athanor Forge de fine-tuning & pipeline RAG
Monad Orchestrateur de micro-agents autonomes

💼 Services ALKYMIA

  • 🖥️ Clusters LLM on-premise clé en main (vos données restent chez vous)
  • 🤖 Agents IA spécialisés : RAG, trading, automation métier
  • 🔌 Intégration MCP / API LLM dans vos workflows
  • 🎓 Formation & conseil : IA souveraine, LLMOps, architecture distribuée

🔧 Stack Technique

Python Docker Swarm CUDA LangChain PostgreSQL Redis FastAPI Whisper Ollama


106 repos · 280 741 lignes Python · 17 hackathons livrés

Pinned Loading

  1. Performance Analyzer - Hardware + Fi... Performance Analyzer - Hardware + FileSystem + Multi-IA Integration - MEGA_ORGANIZED method
    1
    # ═══════════════════════════════════════════════════════════════════════════════
    2
    # PERFORMANCE ANALYZER - Analyse + Maintenance + Optimisation
    3
    # ═══════════════════════════════════════════════════════════════════════════════
    4
    # Intégré au système Multi-IA
    5
    # Scan méthode MEGA_ORGANIZED
  2. APPEL API - UTILISATION MODEL SELECT... APPEL API - UTILISATION MODEL SELECTIONNEE ET OPTIONS ENTREE/SORTIE - Logique complète avec exemples OpenAI et Perplexity
    1
    # APPEL API - UTILISATION MODEL SELECTIONNEE ET OPTIONS ENTREE/SORTIE
    2
    
                  
    3
    ## LOGIQUE D'UTILISATION - GUIDE COMPLET
    4
    
                  
    5
    Ce document presente la logique complete pour utiliser les APIs OpenAI et Perplexity avec selection de modele et gestion des entrees/sorties.
  3. Ponderation Dynamique Multi-IA V2 - ... Ponderation Dynamique Multi-IA V2 - Formules Auto-Optimisees
    1
    # PONDERATION DYNAMIQUE ENRICHIE
    2
    ## Formules Auto-Optimisees Multi-IA
    3
    
                  
    4
    ---
    5
    
                  
  4. Machine Orchestrator - Auto-detect h... Machine Orchestrator - Auto-detect hardware, adaptive allocation, performance optimization
    1
    # ═══════════════════════════════════════════════════════════════════════════════
    2
    # ORCHESTRATOR - Analyse Materiel + Adaptation Automatique
    3
    # ═══════════════════════════════════════════════════════════════════════════════
    4
    # Detecte automatiquement: RAM, GPU, CPU
    5
    # Calcule allocations optimales
  5. Script Test Distribution Massive Mul... Script Test Distribution Massive Multi-IA
    1
    #!/usr/bin/env python3
    2
    """
    3
    TEST DISTRIBUTION MASSIVE MULTI-IA
    4
    Mesure performances reelles et enrichit ponderations
    5
    """
  6. Resultats Tests Massifs Distribution... Resultats Tests Massifs Distribution Multi-IA
    1
    
                  
    2
    # PONDERATION ENRICHIE - RESULTATS TESTS MASSIFS
    3
    # Date: 2025-11-27 19:07
    4
    # Tests: 8 executes
    5