How AI‑Powered “WenZhi” Transforms Job Matching with Baidu’s ERNIE Model

Background and Motivation

During a campus spring recruitment fair, many students found that despite collecting numerous resumes and company brochures, few received job offers. They identified four major pain points: information overload from multiple channels, rapidly changing hiring demands, limited candidate awareness of market needs, and low matching accuracy of traditional keyword‑based search platforms.

Concept and Product Overview

To address these challenges, a team led by PPDE Paddle developer expert Gao Fuzhi created “WenZhi,” an AI‑enabled personal job‑matching system. Users interact with the system through natural‑language, multi‑turn dialogues, providing their resume and job preferences. The app then analyzes the input, matches suitable positions, and offers interview preparation tips.

Technical Architecture

Enterprise Position Database : Positions are vectorized using ERNIE’s semantic‑vector capability, forming a searchable knowledge base.

Matching Agent Construction : Leveraging ERNIE’s function‑calling feature, the system builds distinct agents for resume analysis, demand analysis, recommendation synthesis, and workflow orchestration.

User Input Processing : The user’s resume and job intent are fed into ERNIE via the “dialogue completion” API. Length‑adaptive handling ensures appropriate model selection for varying input sizes.

Result Generation : An agent invokes ERNIE together with Retrieval‑Augmented Generation (RAG) to produce matched positions, salary ranges, location, benefits, and concise interview advice, all within under a minute.

Workflow Enhancements in Version 3.0

The latest release introduces intelligent‑agent orchestration, enabling multiple workflows such as online job matching, offline matching, talent search, and recruitment information queries. The backend dynamically selects the appropriate workflow based on user intent, improving efficiency and flexibility.

Impact and Future Directions

Since deployment in the Paddle Star River community, “WenZhi” has attracted over 8,000 users and helped many students secure suitable positions. Future plans include AI‑driven mock interviews and personalized resume generation, aiming to break information barriers and guide users toward clearer career paths.