Skip to main content
GenAI & LLM SystemsFlagship Project

InterviewXpert — AI-Powered Mock Interview Platform

Dynamic resume-to-question LLM orchestration paired with real-time OpenCV computer vision for candidate behavioral telemetry

Yash Jondhale
Yash Jondhale

AI/ML Engineer & GenAI Systems Architect • Savitribai Phule Pune University (SPPU)

Video Analysis
Real-Time OpenCV Facial Telemetry
Question Generation
Dynamic Resume-to-Question Engine
Latency
<450ms WebSocket Audio/Video Pipeline
Evaluation
Multi-Dimension Candidate Scorecard
01 • Challenge & Context

The Problem Statement

Traditional technical mock interview platforms rely on static question banks that fail to adapt to a candidate's specific background or probe deep technical claims made on their resume. Furthermore, candidates lack actionable feedback on non-verbal delivery: eye contact stability, posture fidgeting, speaking tempo, and emotional composure under pressure.
02 • Engineering Design

Solution Architecture & Implementation

InterviewXpert links a dynamic React.js frontend with an asynchronous FastAPI microservice backend over duplex WebSockets. When a resume PDF is uploaded, a custom parsing pipeline extracts skills, projects, and work experience. The Gemini API dynamically designs a progressive 5-stage interview plan. As the candidate answers through their camera and microphone, an OpenCV pipeline calculates facial landmark alignment, eye gaze drift, and head motion stability while speech-to-text streams answers into a semantic evaluation engine.

03 • Core Innovations

Key Engineering Innovations

Dynamic Contextual Question Branching: If a candidate struggles with distributed systems, the system recursively prompts with diagnostic hints before pivoting to related architectural trade-offs.

Edge-Optimized Gaze Tracking: Built low-compute OpenCV landmark algorithms that track pupil position without requiring heavy GPU acceleration on client hardware.

Holistic Multi-Modal Scorecard: Synthesizes technical code accuracy, communication clarity (WPM and filler words), and emotional confidence into an actionable post-interview breakdown.

04 • Verified Metrics

Performance Benchmarks & Empirical Impact

MetricBefore / BaselineEngineeredNet Impact
Resume-to-Interview TailoringStatic Questions100% DynamicZero Generic Questions
Audio-Video Processing Latency1,200ms380ms68% Latency Reduction
Behavioral Detection AccuracyManual Review91.4% AutomatedInstant Objective Scoring
05 • Technical Breakdown

Technology Stack Architecture

Frontend & Streaming

React.jsWebSocketsWebRTC MediaStreamTailwindCSS

Backend Microservices

PythonFastAPIUvicornAsyncIOPydantic

Computer Vision

OpenCV (cv2)Facial Landmark EstimationEye-Gaze Angle Detection

Generative AI

Google Gemini APIPrompt Engineering ChainingStructured JSON Schema Extraction
Technical Collaboration

Looking for an AI/ML Engineer or GenAI Consultant?

Yash Jondhale delivers production-ready RAG pipelines, FastAPI microservices, and computer vision systems. Available for machine learning roles and enterprise AI advisory.