Project: Reasoning Agents (Azure AI Foundry) - Microsoft Exam Preparation Helper App

[Vivekkosare]

Track

Reasoning Agents (Azure AI Foundry)

Project Name

Microsoft Exam Preparation Helper App

GitHub Username

@Vivekkosare

Repository URL

https://github.com/Vivekkosare/agentsleague.git

Project Description

This is a Microsoft Exam Preparation Helper app, where

1. Student can input their details, background and intention
2. Student can select the certification (fetched from Microsoft Catalog API)
3. Input the no. of weeks student can invest for preparation.
4. Agent 1 sends that info to Agent 2. Agent 2 calls Microsoft learn Catalog API and Microsoft Learn MCP to get information about the exam
5. Agent 3 creates the Study plan
6. Agent 4 creates weekly plan and engagement plan (Agent 2-4 is a subworkflow)
7. The weekly plan will send reminder emails on weekly basis.
8. Here we have humn in the loop that waits for student input.
9. After that Agent 5 creates evaluation plan and creates assessment exam, that student can select answers among the list of answers.
10. Agent 6 creates the guidelines for exam preparation### Demo Video or Screenshots

Demo Video: https://youtu.be/I341l6K-bdc

Primary Programming Language

Python

Key Technologies Used

• Python
• Microsoft Agent Framework
• Microsoft Learn Catalog API
• Microsoft Learn MCP
• Github models (gpt-5, gpt-4o, gpt-4o-mini)
• Azure Container Registry
• Azure Container App Environment
• Azure Container Apps
• Azure Keyvault
• SendGrid for sending weekly email
• Github Actions for deployment

Submission Type

Individual

Team Members

I am an individual contributor in this app. My name is Vivek Kosare

Submission Requirements

• My project meets the track-specific challenge requirements
• My repository includes a comprehensive README.md with setup instructions
• My code does not contain hardcoded API keys or secrets
• I have included demo materials (video or screenshots)
• My project is my own work with proper attribution for any third-party code
• I agree to the Code of Conduct
• I have read and agree to the Disclaimer
• My submission does NOT contain any confidential, proprietary, or sensitive information
• I confirm I have the rights to submit this content and grant the necessary licenses

Quick Setup Summary

1. Clone the repo
2. Go to directory agentsleague\starter-kits\2-reasoning-agents
3. Run "pip install -r requirements.txt"
4. Once installation is done, open terminal or command prompt and run "streamlit run streamlit_app.py"

Technical Highlights

• Implemented Multi Agentic System using Microsoft Agent Framework
• As I am using free tiers of github models, I change models on runtime when selected models runs out of token. So the app doesn't break
• Added a front end UI using the Streamlit framework for python. This made it easy to make it user friendly for testing the app.

Challenges & Learnings

Challenges:

1. I had to finetune the prompts every time , when I want to achieve specific results. This was challenging and learning for me. I got to know that it takes a lot of iterations for building a perfect prompt.
   Learnings:
2. Learnt how to build multi agentic system using Microsoft Agent Framework. As this is an open source library, I didn't have to spend money on Azure Foundry infrastructure
3. Deploying this app as Azure container app (serverless), helps me save money.

Contact Information

linkedin: https://www.linkedin.com/in/vivek-kosare-2b87a660/ and Email: v_kosare@yahoo.com

Country/Region

Sweden