Business Problem

Recruitment teams face significant hurdles that slow hiring and reduce decision quality. Manual resume screening is resource-intensive and time-consuming. HR professionals often struggle to generate and evaluate technical screening questions due to limited subject expertise, making it difficult to assess candidate qualifications fairly. Interview quality varies across recruiters, increasing bias and inconsistency. Additionally, assembling structured post-interview reports demands considerable effort. These challenges collectively extend hiring cycles and impact hiring outcomes.

Technology Solution

RecruitAssist delivers a coordinated, multi-agent AI-powered workflow to automate and enhance key recruitment stages. Key features include:

• Resume Analyzer agent: Uses AI-driven parsing and OCR fallback (AWS Textract) to accurately extract skills and experience from multiple resume formats, including scanned documents. Candidate data is securely stored in AWS S3 and Firestore for compliance and easy access.

• Question Generator agent: Leverages AWS Bedrock's Claude 3 Sonnet large language model to analyze job descriptions and autonomously generate role-specific, recruiter-friendly technical and HR screening questions, aiding objective candidate evaluation even by non-technical HR teams.

• Interview Analyzer agent: Provides real-time transcription of candidate speech (AWS Transcribe Streaming) and live analysis of communication skills, confidence, and topical relevance, delivering subtle, timely cues to recruiters without interrupting interview flow.

• Post-Interview AI synthesis: Aggregates candidate data, job context, and interview insights into structured, objective reports highlighting strengths, risks, and hiring recommendations, reducing bias and aiding consistent decision-making.

• Modular, cloud-native architecture purpose-built for scalable, secure, and adaptable operation across varied team sizes and hiring volumes.

This multi-agent orchestration combined with advanced LLM-based content generation, real-time speech analytics, and intelligent document processing enables faster, smarter, and more objective recruitment.

Technology Stack

Frontend:

• React, TypeScript, Vite,

• Material UI components

• WebSocket client for live audio streaming

Backend:

• Node.js + Express

• AWS SDK v3 (S3, Textract, Transcribe, Bedrock)

• pdf-parse, mammoth

• Firebase Admin (Auth + Firestore)

AI / LLM:

• AWS Bedrock – Anthropic Claude 3 Sonnet

• Agents: ResumeAnalyzer (resume parsing & skill extraction), InterviewAnalyzer (real-time audio analysis), QuestionGenerator (role-specific Q&A generation)

• Pinecone for vector storage

Data Storage:

• Firestore collections (jobs, candidates, interviews, users)

• S3 for original resume files