Changju Ahn(안창주) ChangJu-Ahn

Hi, I'm ChangJu Ahn (안창주) 👋

Solution Engineer @ Microsoft · I build AI that actually ships

I turn real customer requests into shipped, working AI agents — concept → code → deployed, often in minutes.
_{실제 업무 요청을 → 작동하는 AI 에이전트로. 장표가 아니라, 돌아가는 데모와 MVP로 증명합니다.}

🚀 How I work — from real request to running agent

I'm a Solution Engineer at Microsoft who works close to customers — translating a messy, real-world ask into working software, often the same day. I'd rather hand you a demo you can run than a deck you have to imagine.

Real request  ─▶  Understand the business context  ─▶  Build a working AI agent / MVP
     (mail, chat, meeting)         (Work IQ, data, systems)         (demo you can run, not a slide)

🎯 Ship, don't slide. In the AI era, a runnable demo beats a deck. Most of my work ends as a live app or agent a customer can execute themselves.
⚡ Speed as a feature. Ad-image generator in ~10 minutes. An on-site safety-inspection app built faster than instant curry and one-click deployed to Teams.
🧩 Whole-stack, field-first. Front end, agents, MCP tool servers, Azure infra, governance — whatever it takes to make the scenario real end-to-end.
🌏 Bridge global ↔ Korea. I localize Microsoft's newest AI guidance and patterns for Korean enterprises — often on launch day.

"If it doesn't challenge you, it won't change you."

🧠 What I build

Area	Stack & focus
Agentic AI systems	Microsoft Agent Framework · Azure AI Foundry Agent Service · multi-agent orchestration · human-in-the-loop
AI-accelerated delivery	GitHub Copilot · Copilot CLI · Work IQ — from a work item to a running feature
Low-code ⇄ pro-code agents	Copilot Studio · M365 Agent SDK · Direct Line · Model Context Protocol (MCP)
Azure AI platform	Azure OpenAI · AI Foundry · AI Search + Foundry IQ · AI Gateway (APIM) · Vision / Image gen
Enterprise enablement	governance, Content Safety, identity & licensing automation, on-prem process integration

🎬 Featured demos — watch it actually run

Most of my work lives as real video demos on YouTube ▶︎ @changjuahn7582.

From a real work request → an AI agent _{Work IQ + GitHub Copilot CLI, end-to-end. 업무 컨텍스트를 잘 쓰는 개발자.}	Instant ad-image generator with Azure AI _{1 product photo → GPT analysis → Image 2.0. Built in ~10 minutes.}
Multi-agent with Microsoft Agent Framework _{3 expert agents + report agent on Azure AI Foundry, human-in-the-loop (Zava).}	GitHub Enterprise + Copilot onboarding, automated _{Form → Teams approval → Entra ID group → license auto-assign, in 7 min.}
Copilot Studio — Computer Use _{Agents that drive real UIs without APIs. RPA, evolved.}	AI on-site safety-inspection app → Teams _{From idea to one-click Teams deploy — faster than 3-min curry. 🍛}

🛠️ Featured projects

Project	What it is
Microsoft-AI-Agent ⭐	Hands-on labs for building Microsoft AI agents
create_ad_cut	Full-stack OSS "AI Lookbook": product photo → GPT-5.x → GPT-image-2. Azure ACA + SWA + Cosmos + Blob
Azure-AI-Gateway-KR	AI Gateway on Azure API Management — govern models and resources
mock-mes-kr · mock-hr-kr	Throwaway MES / HRIS backends (REST + MCP + SQLite) on Container Apps — instant realism for agent demos
azure_aisearch_workshop	Azure AI Search + the new Foundry IQ knowledge layer, hands-on
azure-superpowers	Deployment layer that takes a Copilot harness from code to a live Azure URL (keyless, single RG)

⚙️ Tech I reach for

Let's build something real. LinkedIn · YouTube · Email

Navigation Menu

Changju Ahn(안창주) ChangJu-Ahn

Achievements