JirayaAi

Your Career Wingman

Comment

Inspiration

When I was in college, I struggled with low confidence, poor communication skills, and had no real interview experience. Getting that experience wasn’t easy — most companies require you to pass an Applicant Tracking System (ATS), and even then, getting a reply felt like winning the lottery. For freshers, many job openings expect 1–3 years of experience, making it extremely hard for someone still learning and growing to even get to the interview stage. I’ve personally faced this frustration. Even when I did land interviews, I failed — not because I didn’t know the answers, but because I lacked preparation and structure. It was painful knowing I could’ve done better with the right guidance. That’s where the idea for JirayaAI came from — a solution to simulate real interviews, build confidence, and prepare people like me. Existing tools on the market are either too expensive, poorly designed, or lack effective prompt engineering. I knew we could do better.

What it does

JirayaAI is an AI-driven web application that offers a full-fledged interview simulation experience. Users can:

  • Choose from multiple interview levels: Basic, Advanced, and Adaptive
  • Answer questions using voice or text
  • Receive instant feedback on their responses
  • View detailed reports and ideal answers after each session
  • Switch between subscription plans and manage their profile
  • Track past interviews and monitor their progress over time

  • All in a smooth, modern, and secure environment — designed to boost confidence and communication skills.

  • Whether you're a beginner or job-hunting pro, JirayaAI helps you prepare better, practice smarter, and build confidence.

How we built it

  • Frontend: Built using Bolt.new, focused on rapid prototyping with a futuristic UI
  • Backend & Auth: Powered by Supabase (migrated from Neon for better compatibility)
  • Voice Agents: Integrated with ElevenLabs for high-quality speech playback
  • AI Models: Used Gemini for generating dynamic, role-specific questions
  • Supplemented with ChatGPT and Claude.ai for fallback content and prompt generation
  • Implemented lazy loading and component-level optimization to keep the experience fast and smooth

Challenges we ran into

  1. Writing effective prompts to generate diverse and relevant interview questions
  2. Avoiding repetitive API calls and unnecessary re-renders
  3. Managing large components and heavy assets with lazy loading
  4. Migrating data from Neon to Supabase to ensure compatibility with ElevenLabs and streamlined authentication
  5. Designing a clean, modern UI that performs well across devices, without lag or clutter

Accomplishments that we're proud of

  • Built a complete full-stack application from scratch
  • Learned to effectively integrate and balance multiple AI tools
  • Understood user experience flows and implemented a seamless, responsive UI
  • Created a project that directly solves a real-life problem we’ve faced ourselves

What we learned

  • The importance of prompt engineering and how deeply it affects AI responses
  • Handling real-world constraints like API limits, state management, and asynchronous behavior
  • Optimizing performance with lazy loading, modular design, and resource handling
  • Crafting a UI that doesn’t just look good, but enhances usability and user confidence
  • The value of iterative testing and listening to user feedback early on

What's next for JirayaAi

  • Multi-language support beyond English and French
  • Community features to share and compare responses
  • Advanced analytics for progress tracking and skill mapping
  • Dark and light theme

Built With

Share this project:

Updates