AI Student Project Manager Agent

AI Student Project Manager Agent — System Architecture
AI Agent Generates a Structured 9-Task Project Workflow
AI Agent Updates Task Status and Tracks Project Progress
Google Cloud Firestore — Persistent Project State and Task Data
GitHub Repository — Project Implementation and Supporting Tools

AI Student Project Manager Agent

Inspiration

Students often have good project ideas but struggle to decide what to do next, track their progress, and manage multiple development tasks. We wanted to create an AI-powered project manager that could act more like an intelligent teammate instead of simply answering questions.

What it does

The AI Student Project Manager Agent converts a student's project idea into a structured development plan. It maintains the project state, identifies the highest-priority pending task, starts tasks, updates task status, marks completed tasks, and reports overall project progress.

For example, the agent can automatically determine the next task, change it to In Progress, and recommend what the student should work on next.

How we built it

The project was developed using Google Agent Development Kit (ADK) with Gemini 3.5 Flash as the AI model.

We created Python tools for:

Creating project tasks
Checking project status
Finding the next best task
Starting the next task
Updating task status
Completing tasks
Calculating project progress

Project state is stored persistently using a structured project data file, and Google Cloud Firestore integration is included for cloud-based project storage.

Agentic AI

The key feature of our system is that the agent does more than generate text. It can take actions through tools and use the current project state to determine what action should happen next.

The workflow is:

Project Idea → Task Planning → State Tracking → Next Task Selection → Task Execution/Status Update → Progress Reporting

This makes the system an agentic workflow rather than a simple chatbot.

What we learned

Through this project, we learned how to:

Build an agent using Google ADK
Connect an AI model with executable tools
Maintain persistent project state
Design task-management workflows
Implement automatic task selection based on priority
Validate task status updates
Integrate Google Cloud Firestore
Test an agent through the ADK development interface

Challenges

One of our main challenges was managing the project's state consistently while allowing the agent to perform actions through tools. We also faced Gemini API quota and availability limitations during development.

Instead of depending on the AI model for every operation, we implemented reliable Python functions for important state-changing operations such as task creation, status updates, completion, and progress calculation.

Impact

Our project aims to make software project management easier for students by providing an AI teammate that helps them understand what to work on next and keeps their project progress organized.

The final workflow successfully tracked 9 out of 9 project tasks, reaching 100% project completion during testing.