The rapid growth of Bangladesh’s job market has created an urgent need for intelligent career guidance and job placement systems that can handle large-scale, unstructured job market data. This paper presents a multi-agent Large Language Model (LLM) framework designed to support career counselling and job recommendation in Bangladesh by integrating web-scraped job market information with machine learning–based salary prediction. The framework includes a web-scraping agent that collects and updates job postings from major employment platforms, a data-cleaning and knowledge-structuring agent that transforms noisy textual job descriptions into structured features such as job category, required skills, and estimated compensation, and a career guidance agent powered by fine-tuned LLMs that delivers personalized recommendations for job seekers based on their academic background, skills, and career aspirations. A salary prediction model is incorporated to enhance transparency and assist candidates in aligning expectations with market conditions. The system was evaluated on a dataset of over 5,000 job postings collected between July and September 2025, demonstrating effective transformation of unstructured data into actionable insights. Combining AI-driven career counselling with data-driven job market intelligence, the proposed framework offers a scalable solution for bridging the gap between job seekers and employers in Bangladesh.
Career Guidance, Large Language Models (LLMs), Multi-Agent Systems, Job Placement, Salary Prediction, Job Market Analytics, Bangladesh.