Feature: AI Mock Interview Scoring & Shareable Result Cards

[MohitBareja16]

Overview

Implement an automated AI scoring pipeline that evaluates a candidate's performance after every coding session. The system will analyse the submitted code and chat transcript to generate a comprehensive score, producing a highly shareable result card (OG-image-ready PNG) to drive viral organic growth on platforms like LinkedIn and Twitter.

Business Value

Tier 1 Differentiator (Viral Hook): No competitor currently auto-scores a session and produces a shareable result card. A candidate posting "NPMChat rated my solution 91/100" serves as free, high-conversion marketing.

Key Requirements

AI Evaluation Pipeline

• Integrate OpenAI GPT-4o to analyse solutions across 4 core dimensions:
  • Correctness (test case pass rate)
  • Time Complexity (Big-O analysis)
  • Code Readability (clean code principles)
  • Communication Quality (analysed via the room's chat transcript)
• Implement an async task queue (bull) to handle the heavy AI processing in the background without blocking the main server thread.

Shareable Artefacts

• Generate a highly polished, OG-image-ready PNG scorecard at a public URL for one-click social sharing. (Using puppeteer or an equivalent image generation library).
• Upload and serve generated images via AWS S3 (or equivalent object storage).
• Generate a detailed, private PDF report exclusively for the interviewer, including per-dimension breakdowns and AI-suggested follow-up questions.

Profile Integration

• Update the User schema to store historical mock interview scores.
• Display the user's historical AI scoring trend/average on their public portfolio page as verifiable evidence of their skills.

Tech Stack

• AI/LLM: OpenAI GPT-4o API
• Image Generation: Puppeteer (or HTML-to-Image equivalent)
• Storage: AWS S3
• Background Jobs: Bull (Redis-backed async queue)