Civy AI - The Future of Interview Coaching

An AI-powered platform using Google Gemini to provide students with personalized roadmaps, resume audits, and mock interviews, bridging the gap between education and employment.

Comment

Inspiration

As an IT undergraduate at SLIIT, I’ve seen many talented peers struggle during job interviews—not due to a lack of technical knowledge, but because of interview anxiety and a lack of practice. Professional coaching is often expensive or inaccessible. I built Civy AI to democratize career coaching, giving every student a personal, 24/7 mentor. With the power of Google Gemini 3, I realized I could simulate a lifelike interviewer that provides the same level of feedback as a human recruiter.

What it does

Civy AI is a comprehensive career ecosystem designed to transform how students prepare for the professional world. Key features include:

  • Interactive Interview Simulation: Gemini 3 acts as a hiring manager, conducting role-specific interviews and asking dynamic follow-up questions based on user responses.
  • Live Performance Scoring: As the user interacts, a lively scoring system monitors and displays real-time metrics for Confidence, Technical Knowledge, Communication, and Accuracy. These scores adapt dynamically based on the user's situation and answers.
  • Comprehensive Post-Interview Analysis: After the session, the AI generates a detailed Interview Report. This includes a breakdown of specific mistakes made, suggested professional insights, and actionable steps to improve for the next session.
  • Civy Proficiency Certificate: Upon completion, users can download a digitally verified certificate featuring a Skill Matrix and Behavioral Insights (like Eye Contact and Vocal Pace). Each certificate is secured with a unique Authentication ID.
  • One-Click Achievement Sharing: Users can instantly publish their accomplishments to LinkedIn, showcasing their interview readiness to potential recruiters with a single click.
  • AI-Powered Resume Audit: Users can upload their CVs to receive instant feedback on formatting, keywords, and content.

How we built it

I developed Civy AI using a modern full-stack architecture:

  • Frontend: Built with React.js and Tailwind CSS for a professional and responsive user experience.
  • Backend: Powered by Node.js and Express to handle real-time data processing and PDF generation.
  • AI Integration: The core intelligence is driven by the Google Gemini 3 API, utilizing advanced system instructions to maintain a consistent "Interviewer" persona.
  • Design: Prototyped in Figma with a focus on student-centric UI/UX.

Challenges we ran into

  • Context Retention: Keeping the AI focused on the specific interview thread required intensive prompt engineering.
  • Real-time Feedback Loop: Engineering the backend to process user answers and update the "lively score" metrics instantly without significant latency.
  • Tone Calibration: Ensuring the AI provided "tough but fair" feedback—mimicking a real-world recruiter without being discouraging.
  • API Quota Management: As a Solo Developer, I had to optimize state management to minimize unnecessary API calls.

Accomplishments that we're proud of

  • Successfully building and deploying a functional AI career platform independently.
  • Developing a verified certification engine that provides students with tangible proof of their growth.
  • Achieving a Natural Conversation Flow where the AI’s responses feel contextual and human-like.

What we learned

  • Prompt Engineering Mastery: How to structure complex prompts to get structured JSON responses for the scoring engine.
  • Multimodal Data Analysis: Learning how AI can analyze vocal pace and accuracy to provide behavioral insights.
  • User-Centric Design: Realizing that visual indicators like live scores significantly reduce user anxiety by providing immediate feedback.

What's next for Civy AI - The Future of Interview Coaching

  • Voice-to-Voice Interaction: Implementing real-time audio so users can speak their answers instead of typing.
  • Multimodal Feedback: Fully utilizing Gemini’s vision capabilities to analyze body language and eye contact via webcam.
  • Industry-Specific Experts: Adding specialized AI personas for fields like Medicine, Finance, and Law.

Built With

Share this project:

Updates