Skip to content

System Context Diagram (C4 Context)

Last Updated: 2026-07-11 Version: v0.2.0

Last Updated: 2026-01-20 Version: v0.2.0 Reference: 5-Layer Architecture


Context Level Overview

The Aries-Serpent/codex system operates within a broader ecosystem of users and external systems:

%%{init: {'accessibility': {'title': 'System Context Diagram<br/>Users and External Systems'}, 'theme': 'base'}}%%

graph TB
 subgraph "External Systems"
 GH[" GitHub<br/>PR Management<br/>Issue Tracking<br/>Workflows"]
 ZD[" Zendesk<br/>Support Tickets<br/>Customer Data<br/>CRM Integration"]
 HF[" Hugging Face<br/>Model Hub<br/>Model Weights<br/>Community Models"]
 MLFLOW[" MLflow<br/>Experiment Tracking<br/>Model Registry<br/>Artifacts"]
 S3[" Cloud Storage<br/>S3/GCS/Azure<br/>Model Storage<br/>Data Backup"]
 BENCH[" Benchmark<br/>Services<br/>Performance Tracking<br/>Leaderboards"]
 end

 subgraph "Aries-Serpent/_codex_ System"
 CORE[" Core Platform<br/>Training<br/>Evaluation<br/>Serving<br/>RAG Pipeline"]
 AGENTS[" Agent System<br/>161 Autonomous Agents<br/>Task Orchestration<br/>Execution Engine"]
 BRAIN[" Cognitive Brain<br/>OODA Loops<br/>Quantum Orchestration<br/>Decision Making"]
 INFRA[" Infrastructure<br/>Config Management<br/>Database<br/>Monitoring<br/>Caching"]
 end

 subgraph "Users"
 DEV[" ML Engineers<br/>Model Development<br/>Experimentation<br/>Training"]
 DS[" Data Scientists<br/>Feature Engineering<br/>Data Analysis<br/>EDA"]
 OPS[" DevOps/SRE<br/>Deployment<br/>Monitoring<br/>Operations"]
 LEAD[" Team Leads<br/>Progress Tracking<br/>Resource Planning<br/>Governance"]
 end

 %% User interactions with system
 DEV -->|"CLI Commands<br/>API Calls"| CORE

 DEV -->|"Configure<br/>Trigger"| AGENTS

 DS -->|"Query<br/>Analyze"| INFRA

 OPS -->|"Deploy<br/>Monitor"| INFRA

 LEAD -->|"Metrics<br/>Reports"| BRAIN

 %% System internal flows
 CORE -->|"Execute<br/>Tasks"| AGENTS

 AGENTS -->|"Request<br/>Context"| BRAIN

 BRAIN -->|"Decision<br/>Feedback"| AGENTS

 CORE -->|"Persist<br/>State"| INFRA

 AGENTS -->|"Log<br/>Metrics"| INFRA

 %% External system integrations
 CORE -.->|"Push Models<br/>Pull Weights"| HF
 AGENTS -.->|"PR/Issue<br/>Actions"| GH
 BRAIN -.->|"Support<br/>Tickets"| ZD
 CORE -.->|"Track<br/>Experiments"| MLFLOW
 INFRA -.->|"Store<br/>Retrieve"| S3
 AGENTS -.->|"Upload<br/>Results"| BENCH

 %% Styling
 style DEV fill:#e0f2fe,stroke:#0284c7,stroke-width:2px,color:#000
 style DS fill:#e0f2fe,stroke:#0284c7,stroke-width:2px,color:#000
 style OPS fill:#e0f2fe,stroke:#0284c7,stroke-width:2px,color:#000
 style LEAD fill:#e0f2fe,stroke:#0284c7,stroke-width:2px,color:#000

 style CORE fill:#dcfce7,stroke:#16a34a,stroke-width:2px,color:#000
 style AGENTS fill:#dcfce7,stroke:#16a34a,stroke-width:2px,color:#000
 style BRAIN fill:#dcfce7,stroke:#16a34a,stroke-width:2px,color:#000
 style INFRA fill:#dcfce7,stroke:#16a34a,stroke-width:2px,color:#000

 style GH fill:#f3e8ff,stroke:#7c3aed,stroke-width:2px,color:#000
 style ZD fill:#f3e8ff,stroke:#7c3aed,stroke-width:2px,color:#000
 style HF fill:#f3e8ff,stroke:#7c3aed,stroke-width:2px,color:#000
 style MLFLOW fill:#f3e8ff,stroke:#7c3aed,stroke-width:2px,color:#000
 style S3 fill:#f3e8ff,stroke:#7c3aed,stroke-width:2px,color:#000
 style BENCH fill:#f3e8ff,stroke:#7c3aed,stroke-width:2px,color:#000

System Responsibilities

Core Platform

  • Training: Model training with hyperparameter tuning
  • Evaluation: Metric computation and benchmark execution
  • Serving: Inference and prediction pipelines
  • RAG: Vector indexing and semantic retrieval

Agent System

  • Task Orchestration: Route and coordinate work
  • Autonomous Execution: Run tasks without human intervention
  • Error Recovery: Self-healing and resilience
  • Feedback Loop: Learn from outcomes

Cognitive Brain

  • OODA Loops: Observe Orient Decide Act
  • Pattern Recognition: Identify recurring issues
  • Decision Making: Choose optimal actions
  • Quantum Orchestration: Coordinate complex workflows

Infrastructure

  • Configuration: Hydra-based setup and secrets
  • Database: Session and checkpoint persistence
  • Monitoring: Metrics, logs, alerts, dashboards
  • Caching: Results, models, embeddings

User Types & Interactions

User Type Primary Interactions Key Workflows
ML Engineers CLI, API, training Develop models, run experiments, deploy
Data Scientists Query tools, analysis Feature engineering, data analysis, EDA
DevOps/SRE Deployment, monitoring Production deployment, system health
Team Leads Dashboards, reports Progress tracking, resource planning

External System Integrations

System Purpose Direction Frequency
GitHub PR automation, workflows Bidirectional Per-commit
Zendesk Support ticket sync Bidirectional Per-ticket
Hugging Face Model sharing Bidirectional Per-release
MLflow Experiment tracking Unidirectional () Per-run
Cloud Storage Model/data persistence Bidirectional On-demand
Benchmarks Performance tracking Unidirectional () Per-release

Key Boundary Crossings

Incoming (Users System)

  1. CLI Commands - User runs codex train --config config.yaml
  2. API Calls - Programmatic model requests: POST /api/predict
  3. GitHub Triggers - PR opened Automated checks
  4. Configuration Changes - Update configs/ System reloads

Outgoing (System External)

  1. Model Upload - Push trained models to Hugging Face
  2. PR Comments - Agent post findings to GitHub PR
  3. Support Tickets - Update customer issues in Zendesk
  4. Metrics Upload - Submit benchmark results
  5. Artifact Storage - Save models/data to cloud storage

Architecture Context

This C4 Context diagram represents Level 1 of the C4 model:

Level 1: System Context (this diagram)

Level 2: Containers (see 5-Layer Architecture)

Level 3: Components (see individual layer docs)

Level 4: Code (see source files)

Next Steps


Related Documentation: - ARCHITECTURE.md - Full architecture documentation - 5-Layer Architecture - Internal system layers - Check the integration documentation in the repository for external system integration details