Inspiration
Traditional mock interviews give feedback after the session is completely over, when it's already too late. We built EchoHire to change that. It is a live, mid-conversation interview evaluation cockpit that tracks and scores technical performance in real-time, helping candidates fix mistakes on the spot.
What it does
EchoHire is a fast Single Page Application (SPA) that tests candidates over a 5-stage engineering question loop:
0-Calibrated Gauge: On load, the confidence meter starts at absolute 0 and moves dynamically based entirely on the quality of your answer.
Smart NLP Matrix: It checks your response word-by-word, immediately differentiating between a weak answer ("ok fine") and an elite technical choice.
Instant Penalty Flag: If an answer is weak, the meter crashes instantly, the status turns to Wrong Answer, and adaptive coaching tips pop up to highlight exactly what coding tools you missed.
Live Graph Tracing: When you use proper engineering terms (Redux Toolkit, useMemo, microservices), the score jumps up, progress bar moves ahead, and a live Chart.js line graph plots your technical growth graph instantly.
How we built it
UI Layout: Built using modern Single Page Architecture (SPA) with Tailwind CSS and Google Material Icons.
Data Telemetry: Integrated Chart.js to drive the dynamic side radial confidence gauge and the interactive progress line chart.
Client-Side AI Simulator: Designed a lightweight, robust semantic token analyzer in JavaScript that maps keyword weights to scoring metrics without server lag.
Challenges we ran into
State Management Sync: Keeping chat data safe while switching between Analysis, History, and Tips tabs without resetting the active interview data. We solved this by creating a localized conversation cache layer.
Context Verification: Fine-tuning the script logic so it don't just scan for buzzwords but verifies if the candidate answered the exact question asked.
Accomplishments that we're proud of
Creating a true, responsive 0-to-Peak dynamic gauge pipeline that reacts flawlessly to real-time text density.
Implementing a smooth multi-view router with absolutely zero framework overhead.
What we learned
We mastered multi-chart asynchronous lifecycles in JavaScript and realized the power of building 100% fail-safe client-side logic that stays functional even when the internet drops.
What's next for EchoHire
Adding real-time WebRTC voice support with Whisper API for live speech-to-text conversion.
Introducing smart generative AI layers to create custom system design follow-up questions based on the candidate’s answers.
Bhai, yeh ek dum short aur high-impact hai. Isko dalo aur submission complete karo!
Log in or sign up for Devpost to join the conversation.