Agentic Architectures Reference
Full implementations located at:
/Users/eddiebelaval/Development/learning/all-agentic-architectures/
Quick Reference Guide
When to Use Which Architecture
Need better output quality? → Start with Reflection (#01)
Need real-time data or external actions? → Use Tool Use (#02) or ReAct (#03)
Need structured, predictable workflow? → Use Planning (#04)
Task requires specialized expertise? → Use Multi-Agent (#05) or Meta-Controller (#11)
Need to prevent errors in production? → Use PEV (#06) or Dry-Run Harness (#14)
Need long-term memory? → Use Episodic + Semantic (#08) or Graph Memory (#12)
Need to explore multiple solution paths? → Use Tree of Thoughts (#09)
Need to test before acting? → Use Mental Loop (#10)
Need robust, less biased conclusions? → Use Ensemble (#13)
Need continuous improvement? → Use RLHF/Self-Improvement (#15)
Need complex emergent behavior? → Use Cellular Automata (#16)
Need self-aware safety? → Use Reflexive Metacognitive (#17)
Architecture Categories
Part 1: Foundational Patterns
Single-agent enhancements for power and structure.
- Reflection - Self-critique and refinement
- Tool Use - External API interaction
- ReAct - Adaptive reasoning + action loop
- Planning - Proactive task decomposition
Part 2: Multi-Agent Collaboration
Specialized teams working together.
- Multi-Agent Systems - Team collaboration
- Blackboard Systems - Shared memory workspace
- Meta-Controller - Smart task routing
- Ensemble - Parallel diverse analysis
Part 3: Advanced Memory & Reasoning
Deep thinking and long-term memory.
- Episodic + Semantic Memory - Human-like memory
- Tree of Thoughts - Multi-path exploration
- Graph World-Model - Interconnected knowledge
Part 4: Safety & Reliability
Production-ready, trustworthy systems.
- PEV (Plan-Execute-Verify) - Self-correcting loops
- Mental Loop (Simulator) - Test before acting
- Dry-Run Harness - Human-in-the-loop safety
- Reflexive Metacognitive - Self-aware limitations
Part 5: Learning & Adaptation
Improvement and emergence over time.
- RLHF (Self-Improvement) - Learn from feedback
- Cellular Automata - Emergent complexity
Integration with ID8Pipeline
Stage Mapping
Stage 2 (Scope Fence):
- Use Planning (#04) to decompose V1 features
- Use Meta-Controller (#11) to route agent capabilities
Stage 3 (Architecture Sketch):
- Use Graph Memory (#12) for entity relationships
- Use Multi-Agent (#05) for component design
Stage 5 (Feature Blocks):
- Use ReAct (#03) for feature implementation
- Use Reflection (#01) for code quality
- Use Tool Use (#02) for external integrations
Stage 7 (Test Coverage):
- Use PEV (#06) for verification loops
- Use Dry-Run Harness (#14) for test safety
Stage 8 (Polish & Harden):
- Use Ensemble (#13) for edge case analysis
- Use Mental Loop (#10) for consequence testing
Stage 11 (Listen & Iterate):
- Use RLHF (#15) for continuous improvement
- Use Episodic + Semantic (#08) for user memory
Implementation Notes
Technology Stack
- Python 3.10+
- LangChain - LLM building blocks
- LangGraph - Stateful agent orchestration
- Jupyter Notebooks - Interactive demonstrations
- Pydantic - Structured data modeling
- Tavily Search - Research tool
- Neo4j - Graph database for memory
- FAISS - Vector store for similarity search
Setup Requirements
- Clone repo to
~/Development/learning/all-agentic-architectures/ - Set up virtual environment
- Install dependencies:
pip install -r requirements.txt - Configure
.envwith API keys (Nebius, LangSmith, Tavily, Neo4j) - Run notebooks in order:
jupyter notebook
Evaluation Pattern
Most notebooks feature LLM-as-a-Judge for quantitative feedback:
- Objective performance metrics
- Rubric-based scoring
- Critical for production AI systems
Learning Path
Beginners: Start with 01 → 04 (Foundational) Intermediate: Move to 05, 07, 11 (Multi-Agent) Advanced: Explore 08, 09, 12 (Memory & Reasoning) Production: Master 06, 10, 14, 17 (Safety & Reliability) Optimization: Study 15, 16 (Learning & Emergence)
Repository: https://github.com/your-username/all-agentic-architectures Last Reviewed: 2026-01-21