My AI & Agentic AI Learning Journey

Building Autonomous AI Agents

💑 CheckYourRelationship.com

  • Built a privacy-first relationship safety platform (React 19, TypeScript, Supabase/PostgreSQL) featuring a 37-question red flag detector, character assessment, and deal-breaker tool with Avg 374 views/week.
  • Integrated Gemini 2.5 Flash AI via a context-aware "Clarity" coach component that injects quiz risk scores and safety flags into dynamic system prompts for personalized responses.
  • End-to-End Ownership: Managed the complete product lifecycle from UX design to backend database schema, prioritizing zero-latency interactions and strict data privacy for sensitive user inputs.

Tech Stack: React 19, TypeScript, Supabase/PostgreSQL, Gemini 2.5 Flash AI

🔗 Visit Website

🤖 Agentic AI Engineering

Comprehensive journey through the Complete Agentic AI Engineering Course, mastering autonomous AI agents and multi-agent systems. The course covers everything from foundations to advanced implementations.

📚 Learning Modules

  • Foundations: Core concepts of Agentic AI, LLM basics, and prompt engineering
  • OpenAI Agents: Building agents with OpenAI SDK, structured outputs, and guardrails
  • CrewAI: Multi-agent collaboration, hierarchical processes, and memory management
  • LangGraph: Workflow automation with graph-based agent systems
  • Autogen: Automated agent generation and team coordination
  • MCP (Model Context Protocol): Advanced agent communication patterns

💡 Generative AI Projects

Collection of practical AI applications showcasing various generative AI techniques and implementations:

  • Marketing Brochure Generator: Automated marketing content creation using OpenAI API
  • Customer Support Bot: Intelligent chatbot with streaming responses
  • RAG Systems: Expert knowledge workers with LangChain and vector databases
  • Code Optimization: AI-powered code analysis and optimization tools
  • Open Source Models: Working with transformers, tokenizers, and model pipelines
  • Budget Tracker: AI-powered financial document analysis

🛠️ Technologies & Frameworks

AI Frameworks: OpenAI SDK, CrewAI, LangChain, LangGraph, Autogen, Hugging Face Transformers

Models: GPT-4, Claude, Gemini, Llama, Mistral, DeepSeek, and various open-source models

Tools: Gradio, Jupyter Notebooks, Vector Databases, RAG Systems, Web Search APIs

Concepts: Prompt Engineering, Function Calling, Structured Outputs, Guardrails, Multi-modal AI

🔄 Agentic Workflow Patterns

  • Prompt Chaining: Sequential agent workflows for complex tasks
  • Routing: Intelligent task distribution across specialized agents
  • Orchestration: Coordinator-worker patterns with parallel execution
  • Evaluation: Agent output validation and quality assurance
  • Research Agents: Deep research with web search and synthesis

🌟 Notable Projects

Deep Research Systems: Multi-agent research platforms with search, analysis, and report generation

Course Instructor Agent: Automated curriculum design, instruction writing, and practice generation

Stock Market Crew: Financial analysis agents using CrewAI framework

Multi-LLM Agents: Integrating multiple models (OpenAI, Claude, Gemini, Bedrock) in single systems

Continuously learning and building at the forefront of Agentic AI technology. Explore my repositories for detailed implementations, code examples, and comprehensive documentation.

🤖 Developer Agents Workflow - Autonomous Code Generation

An intelligent multi-agent system that autonomously handles software development tasks from ticket to pull request. This production-ready workflow automation orchestrates AI agents (coder, QA, PR) to implement features, fix bugs, and create pull requests with zero human intervention.

  • 🤖 Autonomous Development: AI agents (coder, QA, PR) work together to implement features, fix bugs, and create pull requests
  • 🧠 LLM-Driven Intelligence: Uses DeepSeek/Ollama for context-aware code generation and intelligent file selection
  • ✅ Smart QA Validation: LLM decides what tests to run based on patch analysis (syntax, build, Playwright, pytest)
  • 🔄 Self-Healing System: Automatically retries with context-aware error messages when builds fail
  • 📊 Multi-File Patches: Generates Search/Replace blocks for precise code modifications across multiple files
  • 🎯 Semantic Search: Smart file selection using embeddings and LLM-powered relevance scoring
  • 🏗️ LangGraph Orchestration: StateGraph workflow with conditional routing between agents
  • 💾 Supabase Integration: Persistent state management and ticket tracking
  • 🔗 GitHub Integration: Automatic PR creation with detailed descriptions

Tech Stack: Python 3.14, LangGraph, OpenAI/OpenRouter API, Supabase (PostgreSQL), GitHub API, Playwright, Embeddings & Semantic Search

🔗 View on GitHub

© 2022 Judah George.