Skip to content

About

🧠 Self-Improving Multi-Agent System β€” A teaching framework for advanced agentic AI: ReAct loop, Reflexion, multi-agent orchestration, adversarial reasoning, and agent-as-a-judge. With web UI + Netlify deployment.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Latest commit

Β 

History

11 Commits

Folders and files

Repository files navigation

🧠 Self-Improving Multi-Agent System

A teaching framework that demonstrates Reflexion, Multi-Agent Orchestration, Memory, and Tool Use β€” the cutting edge of agentic AI.

What This Teaches

Concept Where to Look
ReAct Reasoning Loop core/agent.py β€” run() method
Reflexion (Self-Improvement) core/reflexion.py
Short + Long-Term Memory core/memory.py
Tool Registration & Calling core/tools.py
Multi-Agent Orchestration agents/orchestrator.py
Adversarial Critic Agent agents/critic.py
Agent-as-a-Judge agents/judge.py
Visual Trace Logging core/trace.py

Architecture

β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚         TASK INPUT                  β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
               β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚   PLANNER AGENT                     β”‚
β”‚   Breaks task into subtasks         β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
               β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚   WORKER AGENTS (parallel)          β”‚
β”‚   Each tackles a subtask            β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
               β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚   CRITIC AGENT (adversarial)        β”‚
β”‚   Finds flaws, challenges outputs   β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
               β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚   REFLEXION LAYER                   β”‚
β”‚   Reflects on failures,             β”‚
β”‚   generates verbal critique,        β”‚
β”‚   updates strategy, retries         β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
               β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚   JUDGE AGENT                       β”‚
β”‚   Synthesizes final answer + scores β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜

Quick Start

# Install dependencies
pip install -r requirements.txt

# Set your API key
export OPENAI_API_KEY="sk-..."

# Run the demo
python examples/demo.py

# Run with a custom task
python examples/demo.py --task "Design a scalable architecture for a real-time chat application"

Design Principles

  1. Readable β€” Every file is small enough to read in one sitting
  2. Honest β€” No magic. Every decision is commented.
  3. Extensible β€” Add a new agent? Subclass BaseAgent. Add a tool? Register it.
  4. Observable β€” Full trace of every thought, action, and reflection.

License

MIT β€” Use it, teach it, break it, learn from it.


Author & Enterprise AI Architecture Hub

Built by Nagarjuna Reddy Kamireddy (Nagarjuna Kamireddy / Nagarjuna K) β€” Lead AI Architect, GenAI & Agentic AI Lead, and Corporate AI Trainer (20+ years across Enterprise Data, SAP HANA & Production AI; AWS Certified Solutions Architect – Professional & NVIDIA-Certified Professional: Agentic AI).

About

🧠 Self-Improving Multi-Agent System β€” A teaching framework for advanced agentic AI: ReAct loop, Reflexion, multi-agent orchestration, adversarial reasoning, and agent-as-a-judge. With web UI + Netlify deployment.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages