Snailclimb/interview-guide

基于 Spring Boot 4.1 + Java 21 + Spring AI 2.0 + PostgreSQL + pgvector + RustFS + Redis，实现简历智能分析、AI模拟面试、知识库RAG检索等核心功能。非常适合作为学习和简历项目，学习门槛低。

What it solves

InterviewGuide provides an end-to-end intelligent interview platform that automates the tedious parts of recruitment and interview preparation. It solves the problem of manual resume screening, the lack of realistic interview practice for candidates, and the difficulty of maintaining a structured, AI-driven question bank based on specific professional knowledge bases.

How it works

The system integrates Large Language Models (LLMs) with a RAG (Retrieval-Augmented Generation) architecture using PostgreSQL with pgvector for knowledge storage. It uses a Spring Boot backend and a React frontend. Key workflows include:

  • Resume Analysis: Asynchronously parses multiple file formats (PDF, DOCX, etc.) via Redis Streams and generates AI-driven evaluation reports.
  • Mock Interviews: Uses "Skill" definitions to generate targeted questions across various domains (Java, Python, AI Agents, etc.) with intelligent follow-up questions and a unified evaluation engine for both text and voice modes.
  • Voice Interaction: Implements real-time streaming dialogue using WebSockets and the Qwen3 model for ASR (Automatic Speech Recognition) and TTS (Text-to-Speech).
  • Knowledge Base: Allows users to upload documents which are chunked and vectorized to enable professional, context-aware Q&A and the generation of specific interview question banks.

Who it’s for

  • Job Seekers: To get AI-powered feedback on their resumes and practice interviews in a realistic environment.
  • HR/Recruiters: To efficiently analyze candidate resumes and assess capabilities at scale.
  • Training Institutions: To manage professional knowledge bases and provide structured interview training for students.

Highlights

  • Multimodal Interviewing: Supports both text-based and real-time voice interviews with low-latency streaming TTS.
  • RAG-Powered Question Banks: Generates interview questions, reference answers, and scoring criteria directly from uploaded knowledge base documents.
  • Skill-Driven Logic: Pre-defined skill sets for 10+ directions (e.g., System Design, AI Agents) to ensure comprehensive coverage and difficulty distribution.
  • Flexible AI Integration: Supports multiple OpenAI-compatible providers including DeepSeek, Kimi, GLM, and LM Studio.
  • Automated Scheduling: Includes an interview arrangement module that uses AI to parse meeting invites from platforms like Zoom or Feishu.