JobSpark

Inspiration

JobSpark was born from the desire to empower South African professionals in their job search. We recognised that while talent abounds, navigating the modern job market can be daunting, especially with the increasing reliance on Applicant Tracking Systems (ATS) and the need for strong interview skills. Our inspiration was to create an intelligent, AI-powered co-pilot that democratises access to career acceleration tools, making professional growth accessible and efficient for everyone. We wanted to build a platform that not only helps users find jobs but also helps them become the best candidates they can be, specifically addressing the unique challenges faced by job seekers in the South African context.

What it does

JobSpark is a comprehensive AI-powered career platform designed to accelerate the job search for South African professionals. It offers a suite of intelligent tools, tailored to overcome local application barriers:

AI-powered CV Generation: Users can build their professional profiles, and our AI crafts compelling, ATS-friendly CVs optimised for specific job descriptions. It can enhance professional summaries and generate detailed experience descriptions, ensuring they resonate with South African employers.
Interview Coaching with Real-time Feedback: Our AI coach conducts mock interviews, asks relevant questions based on role and experience, and provides instant, actionable feedback on content, structure, speech pace, filler words, grammar, clarity, and confidence. This helps build the crucial confidence needed to excel in the local job market.
Personalised Job Matching: The platform connects users with relevant job opportunities, calculating a "match score" based on their CV data (skills and experience) to highlight the most suitable roles in the South African market.
Career Readiness Scoring: JobSpark provides an overall career score, breaking down readiness across profile completeness, CV quality, interview readiness, and market alignment, offering personalised recommendations for improvement, truly empowering users to understand their standing.
Secure Profile Management: Users can securely manage their personal information, work experience, education, and skills, with all data stored and protected via Supabase.

How we built it

JobSpark is built on a modern, scalable tech stack, ensuring a robust and accessible solution for South African users:

Frontend: Next.js 15 with the App Router for a robust and performant web application.
Styling: Tailwind CSS for rapid and consistent UI development, complemented by Framer Motion for smooth animations and interactive elements, enhancing the user experience.
Backend & Database: Supabase serves as our backend, providing a PostgreSQL database for user profiles, experiences, education, skills, generated CVs, and interview sessions. It also handles user authentication, including email/password and Google OAuth, ensuring secure data management for our users.
AI Integration: We leverage Google Gemini AI (specifically gemini-2.0-flash) for all our intelligent features, including CV generation, professional summary enhancement, experience description generation, skill suggestions, interview question generation, and comprehensive interview response analysis, all fine-tuned for South African market nuances.
Text-to-Speech: ElevenLabs API is integrated for natural-sounding AI voice responses during the interview practice, making the experience more immersive and accessible.
Speech-to-Text: We utilise the browser's native Web Speech API for real-time transcription of user responses during mock interviews, enabling immediate analysis.
Deployment: The application is designed for easy deployment on platforms like Vercel or Netlify.

The architecture involves Next.js API routes (src/app/api/) to interact with the AI models and Supabase, ensuring server-side processing for sensitive operations and efficient data handling, crucial for a high-volume platform.

Challenges we ran into

Building JobSpark presented several interesting challenges, many specific to serving the South African market effectively:

AI Prompt Engineering: Crafting effective prompts for Google Gemini to consistently generate high-quality, relevant, and correctly formatted output (especially JSON) for CV content, interview questions, and detailed feedback was an iterative process. Ensuring the AI's responses were tailored to the South African job market and its unique terminologies required careful prompt design and extensive testing.
Real-time Audio Processing: Implementing seamless speech-to-text and text-to-speech for the interview coach, including auto-detection of user pauses to trigger the next question, required careful management of browser APIs and state. Ensuring filler words were accurately captured and analysed was a specific hurdle to provide precise feedback.
Data Modelling & Persistence: Designing a flexible Supabase schema to store diverse user data (profiles, experiences, education, skills, CVs, interview sessions) while maintaining Row Level Security (RLS) for data privacy was crucial, ensuring our users' sensitive information is protected.
Dynamic UI Updates: Reflecting complex data and AI analysis (like career scores and interview insights) in a dynamic and visually appealing manner across different pages, while ensuring data consistency and responsiveness, was challenging but vital for an engaging user experience.
Debugging AI Responses: Sometimes, the AI would return malformed JSON or include extraneous text, requiring robust error handling and post-processing of its output to ensure application stability and reliability.

Accomplishments that we're proud of

We are particularly proud of our achievements in creating a truly impactful solution for South African job seekers:

Seamless AI Integration: The smooth and intuitive integration of advanced AI capabilities directly into the user workflow, making complex tasks like CV writing and interview preparation feel effortless and accessible to all.
Comprehensive Interview Coach: The ability to provide detailed, multi-faceted feedback on interview performance, including speech metrics (pace, filler words, grammar, clarity, confidence), which is a unique and highly valuable feature for building the self-assurance vital for interviews.
Personalised Insights: Delivering actionable, data-driven insights through the Career Score and Dashboard, empowering users to understand their strengths and areas for improvement, and thus strategically improve their job readiness.
User-Centric Design: Creating a clean, modern, and responsive user interface that prioritises ease of use and guides the user through their career journey effectively, ensuring high adoption rates among a diverse user base.
Robust Data Management: Implementing a secure and efficient data storage solution with Supabase, ensuring user data privacy and integrity, which is paramount for our South African users.

What we learned

Throughout this project, we gained significant experience, deepening our understanding of both technical implementation and the specific needs of the South African job market:

Advanced Prompt Engineering: Mastering the art of communicating effectively with large language models to achieve precise and structured outputs, especially in tailoring AI responses to a specific cultural and professional context.
Real-time Web Technologies: Deepening our understanding of Web Speech API, audio processing, and real-time UI updates for interactive applications, crucial for our voice-first features.
Supabase Ecosystem: Leveraging Supabase for authentication, database management, and RLS policies to build secure and scalable applications capable of handling a growing user base.
Full-stack Development with Next.js: Integrating frontend and backend logic within the Next.js framework, including API routes and server-side rendering concepts, for a performant and cohesive platform.
Iterative Development: The importance of continuous testing and refinement, especially when working with AI, to ensure reliability and user satisfaction, leading to a consistently improving product.

What's next for JobSpark

For the future of JobSpark, we envision expanding our impact and utility for South African professionals:

Enhanced Job Matching: Integrating with real-time job boards (e.g., Adzuna, Pnet, Careers24) to provide live, hyper-personalised job recommendations directly within the platform.
Cover Letter Generation: AI-powered generation of customised cover letters for specific job applications, further streamlining the application process.
Career Pathing: AI-driven suggestions for career progression, including recommended skills to acquire and roles to target based on South African market trends and individual user profiles.
Networking Features: Tools to connect users with mentors and industry professionals within South Africa, fostering a supportive community.
Mobile Application: Developing native mobile applications for iOS and Android to provide an even more accessible and convenient experience, catering to the mobile-first nature of many South African job seekers.
Multi-language Support: Expanding support to other official South African languages (e.g., isiZulu, Afrikaans, isiXhosa) to cater to an even broader, more inclusive audience and truly democratise career readiness.