Latest news with #ErranBerger


Forbes
02-06-2025
- Business
- Forbes
LinkedIn Launches AI Job Search Tool For Premium Users
AI is not just reinventing jobs, its reinventing the job search process as LinkedIn launches a new ... More feature for premium users. Job hunting has always involved more friction than most people care to admit. Between vague job descriptions, filter-heavy search tools, and endless scrolling, even well-qualified candidates often struggle to find roles that fit. LinkedIn is betting that AI can fix that. Since May, LinkedIn Premium users have had access to a newly overhauled, AI-powered job search feature that aims to simplify and personalize the way people find work. Built using large language models, the new search tool lets users skip the clunky filters and instead describe what they want in natural language — like 'remote marketing jobs in climate tech that don't require managing a team.' The AI interprets the intent, scans available listings, and returns results tailored to that request. This isn't just a cosmetic update. As Wenjing Zhang, LinkedIn's vice president of engineering, explained in her recent post announcing the release, the goal was to create AI that's fast, scalable and relevant. That last word — relevant — is doing a lot of work in an environment where job seekers are bombarded by scams, outdated listings and ghost recruiters. "One of the biggest changes we've made with our new AI-powered Job Search is to bring more transparency to the job search and recommendations,' said Erran Berger, LinkedIn's vice president of product engineering, in an email reply. 'For example, when a member views a job, they can quickly see how their profile and qualifications match the job requirements. It helps job seekers to be more strategic in their search.' That shift from guesswork to guided strategy is made possible by a 'closed-loop' system that continually improves based on real user feedback. According to Berger, his team reviews both engagement metrics and verbatim user comments weekly, which feed into a set of 'product policies' — a relevance rubric crafted by LinkedIn's jobs experts. These policies help the LLMs grade the quality of search results before they ever reach a user. To ensure those grades aren't biased or broken, every major model goes through pre-launch 'red teaming' — a form of stress testing to identify blind spots. LinkedIn also runs member searches through trust filters to catch potentially harmful or spammy queries. The goal isn't just accuracy. It's integrity. LinkedIn's new AI job search enhancement for premium users relies on natural language processing ... More instead of vague job description guessing. That integrity is visible in other features as well, such as 'Job Match,' which clearly outlines why a particular job was recommended. 'We give job seekers the insights they need — from personalized recommendations to hiring insights — to make smarter, more confident decisions,' Berger wrote. Importantly, users aren't locked into a one-size-fits-all funnel. Berger emphasized that personalization doesn't mean constraint. 'We design our job recommendation systems to empower job seekers while helping them discover new possibilities. Users always retain control over what jobs they search for and apply to," he added. That balance between surprise and control is more than an user experience flourish. It's also a way to empower job seekers. 'This approach not only helps seekers explore new possibilities with confidence, but also enables employers to discover great candidates from non-traditional paths — creating a win-win for both the hirer and the job seeker,' Berger explained. For those who might not know the correct keywords or exact job titles to search for, the AI can serve as a job-hunting sherpa — turning a vague career itch into a concrete list of leads. According to LinkedIn, 40% of job seekers are applying less to job listings or giving up entirely and another 40% are exploring other avenues to try and improve their odds of getting hired. What makes this rollout notable isn't just the tech, but the cross-functional coordination behind it as well as a commitment to real results. 'While we track short-term signals — like job views, saves and applications — we place equal weight on longer-term outcomes, such as whether job posters follow up, interviews progress, and ultimately, whether people get hired,' he noted. That kind of long-game thinking may be what distinguishes this LinkedIn tool from the sea of AI add-ons flooding the job market. 'Our priority is to solve problems for our members and customers. That's why we build our AI-powered products ultimately to add value to our members and customers and help them achieve their goals more effectively and efficiently. More than 90% of those who have tested our AI-powered job features have found them helpful, and early signals from our AI-powered job search are showing similar results,' Berger concluded.


Fast Company
07-05-2025
- Business
- Fast Company
LinkedIn's new AI tools help job seekers find smarter career fits
New AI features from LinkedIn will soon help job seekers find positions that best suit them—without the need for exact keyword matches or specific job titles. LinkedIn's new AI-powered job search interface allows users to express their goals in plain language, says Rohan Rajiv, LinkedIn's head of career products. For example, users can type a phrase like 'business development or partnership roles in video games' and still be matched with relevant positions in the gaming industry, even if job listings don't use those exact terms. Job seekers can also enter more abstract goals like 'using brand marketing skills to cure cancer' to uncover marketing roles at pharmaceutical companies and oncology centers, Rajiv demonstrated in a session with Fast Company. Powered by large language model (LLM) AI, the new search tool interprets the intent behind job seeker queries and job descriptions, making it easier to match people with opportunities in a job market that often frustrates both applicants and employers. 'LLMs are amazing at looking at a job description and extracting semantics out of them and nuance out of them, and inferring things from them,' says Erran Berger, VP of engineering at LinkedIn. Recent survey data from LinkedIn shows that job seekers are applying to more positions than before, increasing the volume of applications that hiring teams must review. 'They say they are spending three to five hours a day sifting through applications, and less than half of these applications meet the required criteria,' Rajiv notes. The enhanced AI search can also surface jobs and employers that applicants may not have previously considered. LinkedIn currently lists more than 15 million job postings. Another new feature, 'job match,' helps users assess their fit for a role before applying. By analyzing both job descriptions and user profiles, the AI identifies how closely someone matches a job's criteria and highlights areas where qualifications are strong or lacking—such as experience with a particular technology. Unlike traditional keyword searches, LLMs are less likely to overlook relevant experience. For example, the AI understands that a web developer likely knows HTML, even if it's not explicitly listed. This insight allows users to refine their profiles with relevant skills or consider roles better aligned with their background, saving time for both candidates and recruiters. Another feature will provide context around whether a position is 'actively hiring,' whether it's being promoted on LinkedIn, and, when available, how long applicants typically wait to hear back. These AI innovations build on tools introduced last year to help recruiters identify strong matches and come roughly 20 years after LinkedIn first began helping users connect with job opportunities. At that time, the company recognized that its vast career and network data could help recruiters discover candidates not actively searching for roles, Berger explains. 'Twenty years ago, LinkedIn transformed the traditional job search by connecting recruiters with passive candidates,' CEO Ryan Roslansky says in a statement to Fast Company. 'And today, with the launch of AI-powered job search, we're not just introducing a new way to find a job, but an entirely new way to discover what's possible.' LinkedIn has steadily expanded its job-matching capabilities, introducing automated job recommendations around 2010 and consolidating job search tools into a dedicated tab on its main platform in 2016. These changes reflect how users often combine job searches with other research activities, such as exploring company profiles. Currently, more than 11,000 LinkedIn users apply for a job every minute. The company expects that its rich history of job and job seeker data will continue to enhance its AI matching tools. 'We're really uniquely positioned to take this rich data that we have—and what's possible with AI now—to build this feature in a way that few, if any, other companies can,' Berger says.