Get-Into

Practice interviews that feel you — emotionally intelligent AI coaching powered by behavioural and emotional patterns

Comment

About the Project

Inspiration

Interviews are not just about what you say, but how you say it. Yet most interview prep tools only evaluate content, ignoring anxiety, confidence, delivery, and emotional signals that heavily influence hiring decisions.

As students and early-career engineers, we’ve all experienced interview stress. Practicing with friends doesn’t replicate real pressure, and existing platforms provide generic, static feedback. We were inspired to build Get-Into.tech after realizing that no tool today combines emotional awareness, fair feedback, and personalized interview coaching.

Our goal was simple: Create an interview practice experience that actually feels like a real interview.

What It Does

Get-Into.tech is an emotionally intelligent AI mock interview platform that simulates real interview pressure while providing empathetic, actionable feedback.

It:

  • Generates company-specific interview questions from real job descriptions
  • Conducts live mock interviews using an AI avatar with voice
  • Tracks biometric and behavioral signals like engagement, posture, and stress in real time
  • Adapts its tone and feedback based on the user’s emotional state
  • Produces a post-session report combining content quality and delivery insights

Instead of saying “your answer was weak,” Get-Into says:

“Your confidence dipped during this question — let’s work on that.”

How We Built It

We designed Get-Into as a multi-agent, AI system:

  • Frontend: React + Vite + TailwindCSS for a smooth single-page experience with webcam and microphone access
  • Backend: Python + Flask for a real-time interview flow
  • AI Reasoning: Gemini API to analyze job descriptions, generate tailored questions, and synthesize feedback
  • Biometrics & Engagement: OpenCV pipeline to detect stress, breathing patterns, and engagement via webcam
  • Voice & Presence: ElevenLabs to power the AI interviewer’s voice and encouragement

Each interview runs as a structured flow where content analysis and emotional signals are evaluated together, not separately.

Challenges We Ran Into

  • Real-time biometric integration: Working with live webcam signals required careful synchronization and fallback handling
  • Latency management: Coordinating voice output, sensing data, and AI responses without breaking immersion
  • Prompt engineering: Teaching the AI to be empathetic, not robotic, took multiple iterations
  • Webcam permissions & browser quirks: Ensuring a smooth user experience under strict browser security rules
  • Scope control: Balancing ambitious features with hackathon time constraints

Accomplishments That We’re Proud Of

  • Built a first-of-its-kind emotionally aware interview coach
  • Successfully combined biometric data + LLM reasoning into a single feedback loop
  • Created a realistic interview presence using voice instead of text
  • Delivered actionable insights that go beyond surface-level scoring
  • Designed a clean, intuitive demo flow judges can experience in under 3 minutes

What We Learned

  • Emotional intelligence dramatically improves AI usefulness
  • AI systems demand strong state management and graceful fallbacks
  • Biometrics add powerful context, but must be interpreted carefully
  • Great UX matters just as much as great models

Most importantly, we learned that AI feels most human when it responds to how people feel, not just what they say.

What’s Next for Get-Into

  • Multi-round interview simulations (phone → behavioral → technical)
  • Adaptive follow-up questions based on detected weaknesses
  • Audio replay with biometric overlays
  • Confidence-building drills and guided breathing exercises
  • Gamified progress tracking and rewards
  • Support for non-native speakers and accessibility features

Our long-term vision is to make high-quality interview coaching accessible, empathetic, and personalized for everyone.

Get-Into isn’t just interview prep. It’s confidence training powered by emotionally intelligent AI.

Built With

Share this project:

Updates