Back to Case Studies

AI-Powered Video Interviewing & Candidate Analysis Platform

AI-Powered Video Interviewing & Candidate Analysis Platform

The Challenge

A forward-thinking HR technology company needed to build a sophisticated digital interviewing platform from the ground up. Their vision was to move beyond simple video recording and create a tool that could intelligently parse resumes, transcribe interviews in real time, and allow recruiters to instantly search through hours of video content to find key information, saving significant time and reducing bias.

Our Solution

ActiveWizards was engaged as the end-to-end engineering partner to bring this vision to life. Our solution was built on three core pillars:

 


High-Level Architecture for the AI-Powered Interviewing Platform

  • Robust Web Platform: We developed a full-stack application using Python/Django for the backend and Vue.js for a dynamic frontend. This included secure user profiles for candidates and interviewers, a real-time video streaming room using Kurento Media Server, and a comprehensive admin interface for management.

  • Intelligent Data Ingestion: To automate candidate onboarding, we integrated Sovren's powerful CV recognition service and built a custom parser for LinkedIn profiles. This allowed candidate data to be structured and ingested automatically from PDFs, DOCX files, and online profiles.

  • AI-Powered Text Analysis: The core innovation was our AI pipeline. We integrated IBM Watson to provide highly accurate, real-time speech-to-text transcription of video interviews. On top of this, we built a custom NLP model capable of not just keyword search, but semantic search that could understand context and identify the beginning of phrases where keywords were used, making the search results dramatically more relevant.

Key Outcomes & Business Impact

  • Drastically Reduced Screening Time: Recruiters could now pinpoint key candidate responses in seconds rather than watching entire interviews, reducing time-to-hire.

  • Automated Data Entry: The platform eliminated manual data entry by automatically parsing resumes and LinkedIn profiles, improving data accuracy.

  • Enhanced Search Capability: The semantic search feature provided a significant competitive advantage, allowing for a much deeper and more intuitive analysis of interview content.

  • Scalable & Secure Platform: The final product was a secure, multi-tenant platform ready to handle a high volume of concurrent video interviews and data processing tasks.

Technology Stack

  • Backend: Python 2.7.x, Django

  • Frontend: Vue.js, HTML5, CSS3, jQuery, SASS

  • Database: PostgreSQL

  • AI & NLP: IBM Watson Speech to Text, Sovren CV Parsing

  • Video Processing: Kurento Media Server, FFmpeg, Node.js