Horizonte I

AI-Powered Career Mentorship & Strategic Roadmap Agent

Comment

Inspiration

The transition from university to the job market is a period filled with uncertainty, career anxiety, and decision fatigue. Many recent graduates and job seekers experience analysis paralysis or feel stuck in positions that do not align with their long-term professional aspirations.

Horizonte I was created to transform career uncertainty into clarity, peace of mind, and actionable strategy. It reframes the user's current professional situation as a deliberate first step toward their dream career by identifying transferable skills and constructing a clear, achievable development path.

What it does

Horizonte I is an interactive AI career mentor that serves as a collaborative learning and growth partner:

  • Profile & Gap Analysis: Evaluates the user's current employment status, existing skill sets, and long-term career goals.
  • Strategic Career Roadmap: Generates a structured 3-stage action plan (Present/Foundations, Mid-Term Transition, and Ultimate Goal).
  • Actionable & Immediate Recommendations: Provides concrete daily learning habits, skill-building guidance, and mindset shifts from day one.
  • Empathetic Guidance: Maintains a supportive, human-centered, and business-oriented tone that alleviates career anxiety and bridges traditional domains with high-value AI roles.

How we built it

The project leverages Google Cloud's Generative AI infrastructure to ensure scalability, reliability, and precision:

  • Agent Brain: Powered by Vertex AI / Gemini 3.5 Pro for natural language processing, transferable skills mapping, and roadmap generation.
  • Orchestration & Logic: Utilizes the Google GenAI SDK in Python for System Prompt management, session memory, and conversation structuring.
  • Backend Architecture: Built with FastAPI to handle web requests and manage data flow efficiently.
  • Cloud Infrastructure & Deployment: Deployed via Google Cloud Run for a secure, scalable web service backend, fully containerized using Docker.

Architecture Overview:

  1. User Interface: Captures user details (current job status, skills, career goals, and concerns).
  2. Google Cloud Run: Processes incoming requests via a Python backend powered by the GenAI SDK.
  3. Vertex AI & Gemini: Executes Horizonte I's System Prompt, assessing skill gaps and structuring the action plan.
  4. Value Delivery: Returns a structured, interactive Career Roadmap and action plan directly to the user.

Challenges we ran into

  • Agent Personality & Tone: Striking the ideal tone in the agent's personality—ensuring it remains professional and strategic without feeling cold or overly technical. We iterated extensively on System Prompt design to deliver clear, actionable responses infused with genuine empathy.
  • Contextual Alignment: Designing prompts that accurately evaluate traditional business backgrounds without oversimplifying the strategic skills required in the modern AI economy.
  • Time Constraints: Balancing backend orchestration, containerization, and prompt refinement within a tight hackathon window.

Accomplishments that we're proud of

  • Successfully connecting traditional business strategy concepts with cutting-edge AI orchestration using Google Cloud.
  • Building a working prototype that converts complex career transitions into clear, actionable steps in seconds.
  • Containerizing the application with Docker and establishing a production-ready Cloud Run architecture.

What we learned

  • Integrating Gemini 3.5 via the GenAI SDK on Google Cloud allows developers to build high-impact AI agents supported by an efficient, production-ready architecture.
  • Validated that advanced AI tools like Gemini can effectively democratize access to high-level strategic career mentorship for non-technical disciplines.

What's next for Horizonte I

  • Integrating real-time job market data via Google Search grounding for live skill demand analysis.
  • Introducing interactive feedback loops for continuous roadmap adjustments based on user progress.

Built With

Share this project:

Updates

Submission history