Zero Queue

ZeroQueue is a serverless AI-powered screener that automates interviews, scores candidates with LLMs, and delivers instant reports using AWS Lambda, Bedrock, and EventBridge.

Comment

🧠 ZeroQueue – AI-Powered Serverless Interview Screener

Inspiration

Recruiters are overwhelmed by the time and bias involved in screening applicants. Scheduling live interviews, reading endless CVs, and maintaining fairness are real bottlenecks. I wanted to create a tool that automates initial interviews in a fair, scalable, and serverless way — using the power of AWS.

What it does

ZeroQueue is an AI-powered, serverless application that:

  • Accepts video/audio interview responses asynchronously
  • Uses AWS Transcribe to convert speech to text
  • Analyzes transcripts using Amazon Bedrock and Comprehend
  • Scores candidates based on tone, sentiment, clarity, and skill match
  • Generates a report and notifies recruiters via email/EventBridge alerts

It automates the screening process, reduces bias, and eliminates the need for real-time scheduling.

How we built it

We built ZeroQueue using a 100% serverless architecture:

  • AWS Lambda: Core logic (transcription, analysis, scoring, reporting)
  • API Gateway: Triggers Lambda functions from the web interface
  • Amazon S3: Stores candidate audio/video submissions
  • AWS Transcribe: Converts voice responses to text
  • Amazon Bedrock (Claude): Evaluates answers for relevance, tone, and alignment
  • AWS Comprehend: Detects sentiment and key phrases
  • EventBridge: Sends structured alerts and triggers follow-up tasks
  • DynamoDB: Stores metadata for job roles and submissions

Frontend was built using HTML/JS and integrates with AWS services via signed URLs and API Gateway.

Challenges we ran into

  • Audio processing flow: Ensuring audio uploads were securely stored and passed to Transcribe.
  • Latency management: Orchestrating async Lambda flows without blocking or duplicate execution.
  • LLM prompt tuning: Crafting effective prompts for Bedrock to produce reliable candidate scoring.
  • Security: Making sure only authorized users could upload or access candidate data.
  • Scalability: Keeping everything stateless and scalable for hundreds of submissions.

Accomplishments that we're proud of

  • Built a complete serverless pipeline using real AWS services
  • Successfully integrated Amazon Bedrock for smart LLM-based analysis
  • Delivered meaningful scoring and recommendations from raw voice data
  • Enabled async interviews with zero queue, zero scheduling, and zero infrastructure

What we learned

  • Deepened knowledge of AWS Lambda triggers and chaining via EventBridge
  • Understood the nuances of working with Bedrock LLMs and prompt engineering
  • Learned best practices in designing stateless, scalable applications
  • Gained practical experience using AWS Comprehend and Transcribe in production

What's next for Zero Queue

  • Add support for follow-up questions based on candidate answers
  • Provide ATS integrations (e.g., Workable, Greenhouse)
  • Multi-language support using Amazon Translate
  • Enable hiring teams to compare multiple candidate scores on a dashboard
  • Expand analysis to include plagiarism detection and soft skills measurement

ZeroQueue is ready to bring smart, fast, and fair screening to any hiring process — with zero bias and zero wait.

Built With

Share this project:

Updates

Submission history