AI Mastermind: Open-Source Coding Agent Workshop
Replaces all commercial and paid ecosystems with fully open-source, free, and self-hostable alternatives. Equip developers to deliver large software projects at an unprecedented pace using community-driven AI coding agents and local model orchestration.
Built for the Agentic Era
Practical, high-impact engineering capabilities. Master the tools and workflows to write multi-file code, automate terminal operations, construct MCP servers, and orchestrate agent swarms in minutes.
Autonomous IDEs & Vibe Coding
Master Cline, Void, and Continue. Build complete 3D interactive games, understand the 8 stages of AI coding, and practice context engineering with local models via Ollama—keeping your private code completely on your machine.
provider: 'ollama',
model: 'deepseek-coder'
});
agent.build3DGame('arena');
Terminal-Native CLI Pair Programming
Harness Aider and OpenCode for git-aware, multi-file terminal refactoring. Code with zero IDE distraction, execute automated tests, and resolve merge conflicts straight from your command line.
Extensibility & MCP Architecture
Connect open coding agents to databases, developer tools, and git platforms using Goose and custom MCP servers. Transform GitHub issues into production-ready pull requests automatically.
Multi-Agent Orchestration & Sandboxes
Deploy swarms of specialized agents using OpenHands and Agent Canvas. Run background tasks, automated code reviews, and live trading workstations inside isolated, self-hosted Docker sandboxes.
3 Comprehensive Tracks. 12 Modules. 100% Free Tools.
A structured 21-hour hands-on curriculum divided into three 7-hour deep-dive tracks. Every tool, agent, model, and runtime used is completely open-source, free, and self-hostable.
The Foundations of Agentic AI & Vibe Coding
The first 7 hours focus on establishing a strong foundation in AI-assisted development and scaffolding complete full-stack applications from scratch using entirely free and open-source coding environments.
Introduction to Open-Source AI Coding Agents
- Hands-On 3D & Game Building: Building a first-person shooter game and a 3D game using Cline, a free, open-source autonomous coding agent that operates seamlessly as a standard VS Code extension.
- The 8 Stages of AI Coding: Understanding the transition from basic prompts to full agent orchestration.
- Vibe Coding Fundamentals: The Missing Manual for Agentic AI Coding and the core practices of flow-state "Vibe Coding."
AI Mechanics & Context Engineering
- How LLMs Work Under the Hood: Tokens, memory, reasoning chains, and deploying local models via Ollama.
- Context Engineering: Mastering system prompts, context windows, and the agents.md file strategy for repo steering.
- Workflow Evolution: The evolution of coding workflows: From YOLO mode to structured autonomous loops.
Open-Source IDE Showdown & Hands-On Building
- Exploring Void: An open-source AI editor built as a fork of VS Code providing a familiar experience with complete data sovereignty.
- Trae AI vs. Continue: Comparing Trae AI (free autonomous agents) with Continue (Bring-Your-Own-Model flexibility and full local execution).
- Comparative App Build: Building a Kanban app across these different open environments to benchmark capabilities.
YOLO Mode & Full-Stack Scaffolding
- Open Infrastructure Setup: Setting up Docker, OpenRouter (for free-tier API routing), and FastAPI for responsible YOLO coding.
- Full-Stack Web App: Building a Next.js website and integrating an interactive AI Digital Twin Chat.
- Autonomous Debugging: Debugging complex features like drag-and-drop mechanics using local open-source models.
Terminal-First Engineering & Autonomous Workflows
Month two transitions from GUI-based assistance to professional-grade CLI engineering. The focus shifts to terminal-native agents and integrating them with open project management tools to automate the software development lifecycle.
Advanced CLI & Terminal-Native Coding
- The Rise of Aider: A terminal-first, open-source AI pair programmer that excels at git-aware multi-file editing and automated repository mapping.
- Zero-Distraction Workflows: Managing full engineering workflows directly from the command line without switching to an external text editor.
- Exploring OpenCode: A high-performance, free open-source coding CLI offering a terminal user interface and support for a vast range of model providers.
Extensibility (MCP, Skills, and Plugins)
- Model Context Protocol (MCP) Deep Dive: MCP servers, hosts, clients, and automated capability discovery within open ecosystems.
- Goose Agent Architecture: Adding custom abilities using Goose, an open-source, extensible agent architecture designed to install dependencies, run tests, and edit files.
- Building Custom Tools: Exposing local file operations, databases, and test runners to agents via MCP servers.
Open Project Management & GitHub Workflows
- Open Issue Tracker Integration: Connecting terminal agents to GitHub Issues or open-source trackers like Plane (replacing Jira).
- Achieving Full Autonomy: Translating an issue ticket into a fully coded Next.js pull request via MCP integrations.
- Disciplined CLI Debugging: Implementing structured debugging strategies, automated test validation, and rollback protocols.
SaaS Platform Build
- Custom Instructions & Skills: Creating tailored prompts, system rules, and skill definitions for local model execution.
- Launch Roadmap: Setting up GitHub tickets, branch protections, and acceptance criteria for a V1 product launch.
- Production SaaS Delivery: Building, testing, and deploying an AI Legal Document Generator with FastAPI and Aider.
Multi-Agent Orchestration & Self-Hosted Sandboxing
The final 7 hours enter the frontier of AI development using advanced open-source platforms. The curriculum covers deploying swarms of coding agents, utilizing self-hosted environments, and orchestrating multi-agent architectures.
Multi-Agent Architectures & Visual Workspaces
- Visual Workspace Transition: Transitioning from single CLIs to OpenHands, an open-source, local-first platform featuring Agent Canvas for managing multiple agent sessions.
- Custom Tools & Plugins: Creating custom tools, marketplace plugins, and auto-trigger review hooks for your agents.
- Agent Roles & Delegations: Orchestrating specialized roles: Architect, Implementation Agent, Tester, and Code Reviewer.
Local Execution & Self-Hosting
- Self-Hosted Agent Runtime: Setting up OpenHands for self-hosting to ensure code execution and agent runtime stay entirely inside your private environment.
- Docker Sandboxing: Running scheduled automations and event-driven runs using open-source Docker sandboxes for maximum safety.
- Zero-Cost Model Routing: Routing prompts to local Ollama clusters or free-tier community APIs with zero telemetry leakage.
Managing Large Codebases & Programmatic Execution
- Enterprise Codebase Best Practices: Best practices for utilizing Aider and OpenHands on large team codebases without exceeding context limits.
- Programmatic Python SDK Control: Driving agents programmatically with Python SDKs available in platforms like OpenHands.
- Automated Agent Loops: Scripting automated issue triage, test generation, and PR validation workflows in Python.
Multi-Agent Capstone Project
- Parallel Cross-Repo Orchestration: Orchestrating tasks using parallel agents across repositories to build complex systems simultaneously.
- Live Trading Workstation Build: Final Capstone: Building and deploying a live Multi-Agent Trader Workstation with real-time market data orchestration.
- 100% Open-Source Runtime: Relying completely on open-source agent frameworks and local inference for private financial processing.
Real Systems Built Across 21 Hours
No toy examples or static tutorials. Every developer builds and deploys 4 real-world systems completely with open-source coding agents and local model orchestration.
3D FPS Browser Game Engine
An interactive browser first-person shooter game and 3D environment scaffolded autonomously using Cline in VS Code. Features physics simulation, procedural arena generation, and collision detection built purely through vibe coding and agent trust loops.
Next.js SaaS Web App + AI Digital Twin
A full-stack Next.js web application equipped with an interactive AI Digital Twin Chat. Features drag-and-drop Kanban mechanics, FastAPI backend endpoints, and Docker containerization using free-tier API routing and local models.
Automated AI Legal Document Generator
An autonomous terminal pipeline driven by Aider CLI that ingests legal contract requirements from GitHub/Plane issues, writes multi-file templating logic, runs automated unit test suites, and submits verified pull requests without manual text editing.
Live Multi-Agent Trader Workstation
A high-frequency financial workstation orchestrated by OpenHands Agent Canvas. Coordinates a swarm of parallel agents (stream ingestion, quantitative analysis, risk constraints, and test execution) running inside isolated, private Docker sandboxes.
Meet Your Instructors
Learn directly from engineers and AI practitioners building production-grade intelligence tools and secure agentic systems daily.
Najad Najeeb
Cybersecurity & AI Specialist
Bachelor of Engineering (B.E.)
Specialist in Enterprise AI deployment, cloud architecture, and cybersecurity at Cisco. Najad bridges the critical gap between advanced LLM engineering, open-source agent orchestration, and secure, responsible operations.
Marjan Muhammed
AI Educator & Technologist
Bachelor of Engineering (B.E.)
An AI generalist focused on democratizing prompt engineering, open-source tooling, and agentic workflows. Marjan translates complex multi-agent architectures into actionable, production-grade developer skills.
Designed for Pure Execution.
A structured learning environment built to bypass theoretical bloat and deliver immediate operational capabilities using 100% free software.
100% Free & Open Source
Zero software licensing costs, zero commercial subscription lock-in. Every agent, IDE, model server, and sandbox is completely free and self-hostable.
21 Hours Pure Hands-On
Engage in active, real-time pair programming, live multi-agent deployments, and terminal refactoring. Absolute zero passive slide reading.
Production Blueprints
Receive immediate repository access to 50+ open MCP templates, battle-tested agents.md prompt blueprints, and isolated Docker sandbox configs.