Inspiration
Australia faces a critical career guidance crisis, particularly affecting the 9.91 million people living in regional, rural, and remote areas (36% of Australia's population according to 2024 ABS data). Our research revealed that 65,000 jobs are advertised online monthly across regional Australia, yet many young Australians struggle to connect their skills with these opportunities.
Government data shows alarming trends: Lower education completion rates in rural areas compared to major cities Reduced participation in higher education among regional students (Australian Institute of Family Studies, 2025) 71% of regional students report feeling disconnected from career opportunities due to geographic isolation
The Australian Connection: We noticed our peers from regional Queensland, Western Australia, and rural Victoria felt particularly lost navigating career choices. Traditional career counseling often focuses on city-centric opportunities, ignoring the massive potential in Australia's regions where 46% of tourism revenue ($107 billion) and 90% of food production originates.
Our inspiration was to democratize career mentorship specifically for Australian students, creating an AI Career Coach that understands the unique landscape of Australian industries, regional opportunities, and the cultural nuances of working across our vast continent - from Perth's mining sector to Tasmania's agriculture, and Darwin's tropical industries.
Please see these article issued by AUS Government: https://www.education.gov.au/australian-curriculum/national-stem-education-resources-toolkit/i-want-know-about-stem-education/which-school-students-need-stem-education/remote-rural-and-regional-students
And many more are available on internet that shows innocent rural childrens are deprived of quality education and faces problem in choosing career.
What it does
The AI Career Coach is a full-stack web application specifically designed for Australian students and permanent residents, addressing the unique career challenges across our vast continent.
Key Features with Australian Focus: Australian Market Analysis: Utilizes Google's Gemini AI enhanced with Australian Bureau of Statistics data to provide career recommendations based on the 65,000 regional jobs advertised monthly and Australia's key industries (mining, agriculture, tourism)
Regional Opportunity Mapping: Special focus on the 36% of Australians living in regional areas, connecting students to opportunities in rural Queensland mining, Victorian agriculture, and Western Australian resources sectors
State-Specific Pathways: Personalized recommendations considering user location, ATAR scores, TAFE/university options, and willingness to relocate within Australia's $500 billion export economy
Australian Institution Integration: Direct links to local universities, TAFE courses, apprenticeships, and government programs like JobActive and regional development incentives
Cultural Relevance: Understanding of Australian workplace culture, Indigenous employment programs, and the unique challenges of our geographic distances
User Journey: Sign-in/Sign-up using firebase.
Holistic Profiling including willingness to work in regional areas
AI-Powered Regional Analysis using local labor market data
Tailored Career Roadmaps with Australian qualifications, salary ranges, and regional opportunities
This directly addresses the career guidance gap affecting 9.91 million Australians in regional areas, ensuring no student is limited by their postcode in accessing Australia's diverse career opportunities.
How we built it
We architected this application using a modern, scalable, and serverless stack, leveraging the power of Google Cloud's ecosystem: Frontend: We built a dynamic and fully responsive user interface using React with Vite for a fast development experience. The UI is styled with Tailwind CSS for a professional, polished look. Backend: Our backend is a lightweight but powerful Python server built with the Flask framework. This handles all our business logic and API requests. Hosting: The React frontend is deployed globally on Firebase Hosting for low latency and automatic SSL. The Python backend is deployed on Cloud Run, a serverless container platform that scales automatically (even down to zero) to handle any amount of traffic cost-effectively. AI Engine: The core intelligence is provided by Google's Gemini 1.5 Flash model. We chose this for its exceptional speed and powerful reasoning capabilities, which are perfect for this real-time application. Database: We used Firestore, a serverless NoSQL database, to securely store user profiles and their generated recommendation sessions, enabling persistent, personalized experiences. Authentication & Security: User sign-up and login are managed by Firebase Authentication. All sensitive information, like our Google AI API Key, is securely stored and accessed using Google Secret Manager, a best practice for cloud security.
Challenges we ran into
Deploying a full-stack application to the cloud for the first time was a significant but rewarding challenge. Our main hurdles were: Cloud Configuration: Navigating the complexities of Google Cloud IAM permissions was a major learning curve. We had to debug issues where our Cloud Run service couldn't communicate with Firestore or Secret Manager, which taught us the importance of correctly assigning roles to service accounts. Serverless "Cold Starts": Our initial backend design was too "heavy" and would time out during the Cloud Run startup process. We re-architected our Python application to use a "lazy initialization" pattern. This professional technique ensures the server starts instantly and only loads heavy libraries (like the AI model) on the first user request, making the application far more stable and efficient in a serverless environment. Prompt Engineering: Getting the AI to consistently return a perfectly structured JSON object without any extra text was a challenge. We iteratively refined our prompt, adding "critical instructions" and safety setting configurations to make the AI's output reliable and prevent our backend from crashing due to parsing errors.
Accomplishments that we're proud of
Building a True Full-Stack, Cloud-Native App: We didn't just build a frontend and a backend; we successfully deployed a complete, scalable, and secure application on Google Cloud, from automated builds to secure secret management. Solving Real-World Cloud Problems: Overcoming the startup timeout issue by implementing lazy initialization was a major breakthrough. It represents a shift from just "making it work" to "making it work right" in a professional cloud environment. Creating a Genuinely Useful Tool: The final application provides real, actionable value. The personalized roadmaps, complete with learning resources and job-prep tips, are something we believe could genuinely help our peers. Harnessing the Power of AI: Successfully engineering a prompt that coaxes a complex, structured, and high-quality response from a generative AI model was a huge accomplishment and central to the project's success.
What we learned
This project was an incredible learning experience. Our key takeaways are: Cloud Architecture is Key: The design of your application matters immensely in a serverless world. Patterns like lazy initialization are not just theoretical concepts; they are essential for building stable and efficient applications. The Importance of Logs: Cloud Run's logs were our single most valuable debugging tool. Learning to read and interpret them was the key to solving every major deployment issue. Security First: Using services like Firebase Authentication and Secret Manager from the start taught us the importance of building secure applications by default, rather than treating security as an afterthought. The Power of Iteration: From refining our UI to engineering the perfect prompt, we learned that the best products are built through continuous testing, debugging, and improvement.
What's next for AI Career Coach
This is just the beginning. We have a clear vision for the future of the AI Career Coach: User Feedback Loop: Implement the "Like/Dislike" feedback buttons to collect data on the quality of AI recommendations. This data, stored in Firestore, can be used to fine-tune the model in the future. Personalized Dashboards: Allow users to track their progress, save interesting careers, and check off completed learning resources, creating a true long-term career development dashboard. Integration with Job Platforms: Connect to job search APIs to show live, relevant job openings in India for each recommended career path. Institutional Partnerships: Develop a version for colleges and universities, using the analytics dashboard to provide them with valuable insights into the career aspirations and skill gaps of their student body, helping them tailor their curriculum to better prepare students for the future.
Log in or sign up for Devpost to join the conversation.