LinkedIn has introduced AI-Powered Career Discovery Tools to enhance the job search experience, recognizing that careers are personal journeys filled with challenges. Research indicates that 60% of professionals seeking new jobs feel frustrated, with 40% either applying less or giving up. The current job search tools are criticized for their one-size-fits-all approach, prompting LinkedIn to reimagine the process using AI for a more human-centered experience.
Tailored Discovery
The new job search experience allows users to articulate their goals in their own words, leading to results that better match their aspirations, including roles they may not have considered. This approach aims to make job-seeking more intuitive and inclusive, moving away from rigid filters and keywords.
Integrating Natural Language
AI Job Search utilizes large language models (LLMs) fine-tuned with insights from LinkedIn’s Economic Graph. This marks the first application of LLMs across LinkedIn’s search and recommendation systems, enhancing the understanding of user intent and natural language queries. A cross-encoder LLM is employed to score and rank jobs accurately, while an LLM distillation technique enables real-time job ranking efficiently.
Robust Evaluation
To ensure job relevancy, LinkedIn employs structured evaluation guidelines known as relevance rubrics, developed by job experts. These rubrics define what constitutes "relevant" in relation to specific queries and jobs. LLMs are then used to apply these guidelines, helping to surface jobs that align closely with members' goals and skills.
AI Job Search represents a significant step toward providing personalized support for job seekers, with ongoing testing and exploration of AI capabilities to create even more customized experiences in the future.
