MockMaster AI: Interview Simulation & Feedback Platform

From Practice to Placement — Master Interviews with AI.

Comment

Inspiration

The idea behind MockMaster AI came from a common problem faced by students and job seekers — lack of real interview practice and meaningful feedback. While many platforms provide questions, very few simulate a real interview environment with behavioral analysis.

During the hackathon, when we were required to use MeDo, we saw an opportunity to go beyond a basic chatbot and build a complete AI-driven interview ecosystem that evaluates not just answers, but also confidence, communication, and behavior.

What it does

MockMaster AI is a smart interview preparation platform that:

Conducts AI-powered mock interviews (HR, Technical, Behavioral) Uses voice + camera analysis to assess: Confidence Eye contact Speaking speed Posture Generates role-specific questions dynamically Provides: Question-wise feedback Strengths & weaknesses Ideal answers Creates a personalized interview roadmap Tracks user performance over time via dashboard analytics

It essentially replicates a real interview experience with an AI interviewer + evaluator combined.

How we built it

I built the platform using MeDo as the core AI engine, combined with:

Frontend: Interactive UI for interview simulation, dashboard, and feedback Backend: Session management, user data, and interview history storage AI Modules: Question generation engine (role + type based) NLP-based answer evaluation Confidence scoring system (speech + behavior signals) Voice Processing: Speech-to-text for real-time answer capture Camera Analysis: Face detection + posture + engagement tracking Data Layer: Stores interview results, scores, and progress tracking

The system architecture closely follows a modular design:

Interview Engine Evaluation Engine Behavior Analysis Module Roadmap Generator

Challenges we ran into

Real-time AI feedback Generating instant, meaningful suggestions during answers was complex Voice + Camera Integration Synchronizing speech input with behavioral analysis in real time Confidence Scoring Logic Designing a fair scoring system combining multiple parameters: Speech speed Pauses Eye contact Answer quality Dynamic Question Generation Ensuring questions feel realistic and not repetitive Hackathon Constraint (MeDo) Adapting all AI workflows strictly within MeDo ecosystem

Accomplishments that we're proud of

Built a fully functional end-to-end AI interview simulator Successfully implemented: Real-time speech-to-text + feedback Behavior-based confidence scoring Created a structured feedback system (strengths, weaknesses, improvements) Designed a personalized roadmap generator Delivered a SaaS-like polished experience within hackathon time

What we learned

How to design human-like AI interactions, not just Q&A systems Importance of UX in AI products (feedback clarity matters more than accuracy alone) Combining multiple AI signals (voice + behavior + NLP) is powerful but complex Working within constraints (like MeDo) improves problem-solving Real-world systems require balancing accuracy, performance, and usability

What's next for MockMaster AI: Interview Simulation & Feedback Platform

Advanced emotion detection (facial expressions, stress analysis) More accurate AI evaluation using fine-tuned models Company-specific interview simulations (Google, Amazon, etc.) Advanced analytics dashboard (deep performance insights) Multi-language interview support Group discussion (GD) simulation module Mobile app version Integration with real job preparation platforms

Built With

  • api
  • context
  • css
  • customalgorithms
  • deno
  • enums
  • localstorage
  • lucide
  • mediadevices
  • react
  • recharts
  • router
  • shadcn/ui
  • speech
  • supabase
  • tables:
  • tailwind
  • typescript
  • vite
  • web
Share this project:

Updates

Submission history