Skip to main content

Launch Post Draft

Use these drafts as a starting point for the public release. Keep the tone practical and developer-focused.

Short LinkedIn Post​

I just open-sourced Agentic Engineering Playbook: a build-first learning platform for developers who want to go beyond AI demos and learn how production-shaped agent systems are actually assembled.

It includes 6 runnable AI systems covering:

  • LLM gateway patterns
  • RAG and retrieval evals
  • workflow orchestration and approvals
  • MCP-style tool layers
  • Angular copilot UX
  • safe QA browser agents

Live docs: https://ankitparekh007.github.io/Agentic-Engineering-Playbook/

Repo: https://github.com/AnkitParekh007/Agentic-Engineering-Playbook

Longer LinkedIn Post​

I’ve open-sourced Agentic Engineering Playbook, a practical roadmap for developers who want to learn AI agent engineering by building real systems instead of collecting notes.

The repo is structured around 6 runnable, production-shaped starter projects:

  1. AI Provider Gateway
  2. Enterprise RAG Copilot
  3. Agent Workflow Orchestrator
  4. MCP Enterprise Toolkit
  5. Angular Agentic Copilot
  6. QA Browser Agent

The goal was to make the learning path cumulative:

  • start with model access and structured outputs
  • move into retrieval and grounded answers
  • add orchestration, approvals, and tracing
  • connect tool layers and UI
  • finish with safe browser automation

Everything is documented in a Docusaurus site and backed by starter implementations, local smoke checks, and evals where appropriate.

This is not a finished enterprise platform. It is a set of production-shaped starter projects meant to teach architecture, system boundaries, and upgrade paths.

Live docs: https://ankitparekh007.github.io/Agentic-Engineering-Playbook/

Repo: https://github.com/AnkitParekh007/Agentic-Engineering-Playbook

If you’re building agents, copilots, RAG systems, MCP integrations, or operator-facing AI products, I’d appreciate feedback.

GitHub Launch Announcement​

Launching Agentic Engineering Playbook: an open-source learning platform for developers who want to learn agentic AI engineering by building.

What’s inside:

  • a polished docs site
  • 6 runnable AI systems
  • LLM gateway, RAG, orchestration, MCP-style tools, Angular copilot UI, and QA browser agent examples
  • local evals, smoke checks, and CI coverage

This is aimed at engineers who want a more concrete path from “LLM API demo” to “system with approvals, traces, citations, tooling, and UI.”

Live docs: https://ankitparekh007.github.io/Agentic-Engineering-Playbook/

Twitter / X​

Open-sourced: Agentic Engineering Playbook

6 runnable AI systems for learning by building:

  • LLM Gateway
  • Enterprise RAG Copilot
  • Agent Orchestrator
  • MCP-style Toolkit
  • Angular Copilot UI
  • QA Browser Agent

Docs + code + evals + CI.

Live: https://ankitparekh007.github.io/Agentic-Engineering-Playbook/

Reddit / Hacker News Style​

I open-sourced Agentic Engineering Playbook, a practical curriculum for learning agentic AI engineering through runnable starter projects.

The repo focuses on production-shaped system patterns rather than one-off prompt demos. It currently includes 6 small but runnable systems:

  • AI Provider Gateway
  • Enterprise RAG Copilot
  • Agent Workflow Orchestrator
  • MCP Enterprise Toolkit
  • Angular Agentic Copilot
  • QA Browser Agent

The docs are intended to connect the layers, and the projects are meant to be upgraded over time rather than treated as finished products.

I’d be interested in feedback from people building RAG systems, internal copilots, orchestration runtimes, or tool-using agents.