A comprehensive collection of AI agent implementations and experiments using various frameworks and technologies. This project serves as a learning resource for building intelligent agent systems.
This repository contains multiple sub-projects, each exploring different AI agent frameworks and approaches:
An AI agent service built with FastAPI and the Strands Agents SDK, featuring:
- FastAPI web framework for the agent service
- Strands Agents framework with Ollama and OpenAI-compatible integration
- Docker Compose for service orchestration
- Ollama for model hosting and serving
- llama.cpp for CPU-only inference backend
Quick Start:
cd AWS-Strands
make setupA Python project built with LangGraph and Pydantic AI for creating intelligent agent workflows:
- LangGraph for building stateful, multi-agent applications
- Pydantic AI for type-safe AI agent development
- Structured Logging with structlog
- Configuration Management with Pydantic Settings
- Async Support for efficient execution
- Comprehensive Testing with pytest
Quick Start:
cd langgraph
uv syncA Go project built with the official Google ADK for Go, featuring:
- Google ADK for Go - Official Agent Development Kit from Google
- Idiomatic Go - Natural Go patterns and idioms
- Rich Tool Ecosystem - Pre-built tools, custom functions, and integrations
- Modular Multi-Agent Systems - Scalable applications with specialized agents
- Cloud-Native Deployment - Easy containerization and deployment
- Configuration Management - Environment-based configuration with envconfig
- Structured Logging - Production-ready logging with Uber Zap
Quick Start:
cd google-adk
go mod tidy
go run main.go- Python 3.10+
- Go 1.21+ (for Google ADK)
- Docker and Docker Compose (for AWS-Strands)
- UV package manager (for LangGraph)
- Git for version control
The root directory contains:
pyproject.toml- Project-wide tool configurations (Black, isort, Ruff, MyPy, pytest).agents/- AI agent rules and configurations.venv/- Python virtual environment.vscode/- VS Code workspace settings
The project uses the following development tools with consistent configurations:
- Black - Code formatting (line length: 120)
- isort - Import sorting (compatible with Black)
- Ruff - Fast linting (selects E, W, F, I, B, C4, UP rules)
- MyPy - Static type checking (strict mode)
- pytest - Testing framework
-
Clone the repository:
git clone <repository-url> cd learning-ai-agents
-
Set up the virtual environment:
python -m venv .venv source .venv/bin/activate # On Windows: .venv\Scripts\activate
-
Explore the sub-projects:
- Navigate to
AWS-Strands/for FastAPI-based agent service - Navigate to
langgraph/for LangGraph-based agent workflows - Navigate to
google-adk/for Google ADK Go agent project
- Navigate to
This repository serves as a learning resource for:
- Understanding different AI agent frameworks
- Experimenting with various LLM integration patterns
- Building production-ready agent services
- Learning best practices in AI agent development
This project is for educational purposes. Please refer to individual sub-projects for their specific licenses.