CareerPath.AI

CareerPath.AI is a multi-agent career counseling system using Google’s ADK and Gemini to recommend personalized career paths with explanations and roadmaps.

Comment

Inspiration

Many students struggle to choose the right career path due to lack of guidance, confusion, or being unaware of their strengths. We’ve personally seen friends who are unsure about what to do after graduation, even though they have specific interests and skills. This inspired us to build CareerPath.AI, a personalized career counseling assistant that acts like a virtual mentor, giving tailored recommendations using the power of AI and multi-agent systems.

What it does

CareerPath.AI takes your personal inputs—like your interests, strengths, GPA, and name—and guides you through: A career recommendation tailored to your profile An explanation about why it suits you A step-by-step roadmap for reaching that career All of this is handled by smart agents working in the background, powered by Google’s Agent Development Kit (ADK) and Gemini models.

How we built it

We used: Python for building the backend logic Google ADK to create and manage the agents Gemini 1.5 Pro for generating content Streamlit for the UI GitHub for version control Firebase for potential future use (data storage) Multi-Agent Architecture: InputCollectorAgent: gathers user data CareerRecommenderAgent: suggests a career ExplainerAgent: explains the career fit RoadmapAgent: gives a practical roadmap OrchestratorAgent: coordinates the entire flow Each agent runs independently but works together as a system.

Challenges we ran into

          Misconfigured model names (gemini-pro instead of gemini-1.5-pro) caused runtime issues
          GitHub blocked pushes due to accidentally committed secrets
          Rate limits and quota errors from the Gemini API while testing
          Initial confusion around ADK structure and orchestrating multiple agents
          Streamlit integration with agents required careful handling of asynchronous logic

Accomplishments that we're proud of

           Successfully built a working multi-agent system using ADK
          Deployed the entire project with a clean and responsive UI
          Solved critical API errors and GitHub security blocks during development
          Built a career counselor that gives useful, human-like responses
           Learned to integrate real-world LLMs into a full-stack solution

What we learned

          How to use Google’s Agent Development Kit (ADK)
          Proper API key management and .gitignore best practices
          How to handle rate limits, model configuration, and orchestrate multiple agents
          Using Gemini models efficiently for real-time recommendations
          Deploying an AI-based app using Streamlit
         Writing clean, modular Python code for agent collaboration

What's next for CareerPath.AI

       Add more agents, such as:
                            A MentorAgent to simulate real-life mentorship
                            A CollegeSuggestorAgent based on career choice
      Store session history using Firebase or Firestore
      Allow users to download their career plan as a PDF
      Add voice input and response for accessibility
      Deploy on Google Cloud for better scalability

Built With

Share this project:

Updates

Submission history