LinkedIn Enhances Job Search with AI-Powered Natural Language Queries

LinkedIn recently unveiled a major AI overhaul of its job search functionality, aiming to enhance the user experience by leveraging large language models (LLMs) to understand and deliver more accurate job matches. The new search feature, available to all LinkedIn users as of June 16, 2025, allows members to describe their job goals in natural language, rather than relying solely on precise keyword queries. This update addresses a significant pain point: users often received irrelevant job listings due to the limitations of keyword-based searches. Erran Berger, vice president of product development at LinkedIn, emphasized the importance of this shift, stating, "This new search experience lets members describe their goals in their own words and get results that truly reflect what they’re looking for. It is the first step in a larger journey to make job-seeking more intuitive, inclusive, and empowering for everyone." Previously, LinkedIn's job search function depended on fixed, taxonomy-based methods and older LLMs, which lacked the sophistication to deeply understand user queries. For instance, a search for "reporter" might yield roles in media journalism alongside court reporting positions, despite the different skill sets required. Wenjing Zhang, vice president of engineering at LinkedIn, highlighted this issue, explaining, "When we’re using keywords, we’re essentially looking at a keyword and trying to find the exact match. And sometimes in the job description, the job description may say reporter, but they’re not really a reporter; we still retrieve that information, which is not ideal for the candidate." To address these challenges, LinkedIn overhauled its search function by integrating modern, fine-tuned LLMs, which offer enhanced natural language processing (NLP) capabilities. These models better interpret the nuances of user queries, allowing for more relevant job listings. Users can now enter complex queries like, "Find software engineering jobs in Silicon Valley that were posted recently," and receive tailored results. However, the integration of LLMs is not without its challenges, particularly the high computational costs associated with running these models. LinkedIn mitigated this issue through a process called distillation, where they split the LLM into two parts: one for data and information retrieval, and the other for ranking results. By using a teacher model to rank the initial query against the job listings, LinkedIn was able to align both retrieval and ranking phases, ensuring that the most relevant jobs appear at the top of the search results. This approach also simplified LinkedIn's job search pipeline, reducing the number of stages from nine to a more streamlined process. Wenjing Zhang explained, "To do this, we use a common technique of multi-objective optimization. It's crucial that retrieval ranks documents using the same MOO that the ranking stage uses, keeping the retrieval process simple without overburdening AI developer productivity." An additional innovation is the query engine that generates customized suggestions for users. This engine helps refine search terms and provides more precise recommendations, further enhancing the job search experience. The impact of this AI-driven job search is significant, as it aligns with broader trends in enterprise search. Google predicts that 2025 will be the year when enterprise search becomes more powerful due to advanced models. Similar technologies, such as Cohere's Rerank 3.5, are breaking down language barriers within organizations, making it easier for employees to access and analyze internal data sources. LinkedIn's efforts in AI extend beyond job search. In October 2024, the company launched an AI assistant to aid recruiters in finding top talent. This tool uses LLMs to analyze resumes, recommend candidates, and streamline the hiring process. Deepak Agarwal, LinkedIn's Chief AI Officer, will elaborate on the company's AI initiatives, including the development and scaling of its Hiring Assistant, during the upcoming VB Transform conference in San Francisco in July 2025. Industry insiders are optimistic about LinkedIn's AI overhaul. They believe that by implementing LLM distillation and custom query engines, LinkedIn has set a new standard for job search platforms. The enhancements not only improve user satisfaction but also position LinkedIn as a leader in AI-driven professional networking tools. Additionally, the reduction in computational costs while maintaining performance efficiency underscores LinkedIn’s commitment to sustainable technological advancements. Company Profile: LinkedIn is a professional social network owned by Microsoft, connecting millions of professionals worldwide. It offers a range of services including job listings, resume building, and networking. The company has been at the forefront of integrating AI into its platform, continually innovating to enhance user experiences and operational efficiency.