Career Guidance Agent (Azure AI Projects + Streamlit)

This project is a small demo showing how to use Azure AI Projects with a custom Agent and expose it through a simple Streamlit UI.

Users type a career-related prompt (e.g., “Guide me for Python backend developer”) and the Azure Agent returns a structured, helpful response in Markdown.

---

Demo

---

How I installed from scratch

pip install --pre "azure-ai-projects>=2.0.0b1"
pip install azure-identity
pip install openai
pip install python-dotenv
pip install streamlit=1.45

Since I am using agent built from Azure AI Foundary by so you will need to create agent there by going to Microsoft Foundry

and then from your local you need to configure Azure login using

az login

If you don't have azure cli installed you can do it using brew. Now being a developer don't ask me how to do that, not everything can be vibe coded, just search it on google and do that.

---

🔧 How It Works

The project is split into two parts:

1. main.py

   • Connects to Azure AI Projects
   • Creates or references a career-guidance agent
   • Sends the user prompt
   • Prints the agent's response in Markdown
   • Used internally by Streamlit (run as a subprocess)

2. app.py

   • Streamlit interface
   • Textbox for user prompt
   • A Submit button
   • Calls main.py and displays its output nicely

---

🚀 Getting Started

1. Clone the Repo

git clone https://github.com/ayushdoesdev/career-guidance-agent.git
cd career-guidance-agent

2. Create a Virtual Environment

python -m venv venv
source venv/bin/activate   # macOS/Linux

3. Install Dependencies

pip install -r requirements.txt

(Include Streamlit + Azure SDK packages in your requirements.txt.)

4. Set Up Environment Variables

Create a .env file:

AZURE_EXISTING_AGENT_ID="career-guidance-agent:1"
AZURE_ENV_NAME="..."
AZURE_LOCATION="..."
AZURE_SUBSCRIPTION_ID="..."
AZURE_EXISTING_AIPROJECT_ENDPOINT="..."
AZURE_EXISTING_AIPROJECT_RESOURCE_ID="..."
AZURE_EXISTING_RESOURCE_ID="..."

Make sure you have authenticated using az login or have set service principal credentials.

---

▶️ Running the App

Start Streamlit:

streamlit run app.py

Enter a prompt, click Submit, and the response from your Azure Agent will be displayed.

---

🧩 Project Structure

.
├── main.py        # Handles Azure agent communication and prints response
├── app.py         # Streamlit UI
├── .env           # Environment variables (not committed)
└── requirements.txt

---

✨ Features

• Simple, clean UI for user prompts
• Markdown response rendering
• Uses Azure AI Projects agent
• Modular design (UI + agent code separated)

---

📌 Notes

• Streamlit calls main.py as a subprocess, so the output from main.py must be clean Markdown.
• Agent creation happens inside main.py. If you already have an existing agent, set AZURE_EXISTING_AGENT_ID to avoid creating new versions.
• Make sure you run Streamlit from the same environment that has the Azure SDK installed.