🎉 We've just raised $9.5M Seed round. Read more about our plan ->

Climatebase Increases Job Applications by 50% using Superlinked

Climatebase is the world’s leading hiring platform for climate careers, featuring more than 40,000 job listings and serving over 1 million people each year. The Climatebase Fellowship, launched in 2022, has already accelerated the careers of more than 3,200 Fellows and will welcome its next cohort in the fall. Its “Climatebase Weekly” newsletter keeps 120,000 readers up to date on important climate news and new opportunities. Climatebase also organizes SF Climate Week, California’s premier climate-solutions summit, which attracts 25,000 attendees across more than 500 events created in partnership with over 1,000 organizations.

Climatebase must surface highly relevant job recommendations. Its previous search-based system, powered by Algolia, produced only about a 1% view-to-application rate with many users unhappy with the relevance of their keyword-based search results. Climatebase were looking for a smarter, personalised recommendation engine to lift engagement and match quality.

‍

Climatebase’s Challenge

In the competitive hiring landscape, especially for mission-driven climate careers, personalization is key. Climatebase’s users range from software engineers to policy experts, and presenting the right opportunities to each user is critical for engagement. Under the Algolia-powered system, keyword-based search results were not keeping pace with user expectations. The search often surfaced roles too generic or outside a user’s interest area, leading to a 0.7% job dislike rate and user feedback that many recommendations were off-target (over two-thirds of Algolia feed dislikes were due to the job’s role type not fitting the user). The one-size-fits-all, keyword-based approach struggled to capture the nuanced preferences of the audience. Climatebase therefore needed an automated engine that could infer user intent beyond keywords, increase saves and applications, and minimize irrelevant suggestions. Superlinked provides that leap. By leveraging vector embeddings, it surfaces semantically relevant jobs even in cold-start scenarios, matching each job seeker to promising opportunities before any interaction data exists.

‍

The Superlinked Solution

In May 2025, Climatebase replaced Algolia with Superlinked’s AI-powered recommendation platform to drive its personalized job feeds (including “jobs_directory_recommended_jobs”, “dashboard_recommended”, and “similar_jobs”). Superlinked’s solution uses advanced vector search technology to match candidates with jobs based on semantic relevance, rather than simple keyword overlap. By encoding rich information about job postings (titles, descriptions, required skills, etc.) and user profiles (job title, work experience) into high-dimensional vectors, Superlinked can surface climate jobs that align much more closely with each user’s background and interests. The system updates in real time with user behavior – as candidates browse, save, or apply to jobs, the recommendations adjust to reflect their demonstrated preferences.

Under the hood, Superlinked uses a mixture-of-encoders architecture that combines specialised language and skill encoders with metadata-aware embeddings. This approach blends deep semantic understanding with structured fields such as location, seniority and contract type, enabling precise filtering without sacrificing relevance. The result is richer, context-aware vectors that take into account the job type as part of the search, without the need for user filters.

This integration was seamless: Superlinked’s API was plugged into Climatebase’s existing infrastructure, immediately powering the Recommended Jobs sections for logged-in users. (Notably, these Superlinked-powered feeds serve mostly signed-in users – who generally exhibit higher engagement – but even accounting for that, the uplift in performance was striking.) With Superlinked handling the heavy lifting of relevancy matching and real-time personalization, Climatebase’s team could focus on its mission to connect talent with climate solutions, confident that an intelligent system was curating the job feeds for each user.

‍

‍

Results

The switch to Superlinked brought substantial improvements across Climatebase’s key metrics. In just the first week after deployment, the benefits of more relevant, AI-driven recommendations became clear

‍

‍

  • Higher conversion of views to applications: Superlinked delivered a 1.57% view-to-application conversion rate, compared to 1.04% under Algolia a 50.% + increase in conversion efficiency. Users were significantly more likely to apply for a job after viewing it in a Superlinked-powered feed than they were from the Algolia feed.
  • Increased user engagement (bookmarks): The improved relevance also led to users saving (bookmarking) more jobs per search. The Superlinked-driven feeds yielded only 38% fewer total job bookmarks than Algolia’s feed, even though Algolia served over 3Ă— more queries. This means that Superlinked had double the bookmarking rate per query than Alogilia, a clear signal that users found its matches far more relevant and engaging.
  • Reduction in role-mismatch dislikes (location fine-tuning still in progress):
    User feedback shows a 50% increase in satisfaction with the job types surfaced by Superlinked’s search system. By applying Superlinked’s mixture of encoders and semantic role matching, the new search cut role-mismatch complaints roughly in half compared with the earlier keyword approach.

‍

The Superlinked Advantage

By adopting Superlinked’s AI-powered recommendation system, Climatebase unlocked several key advantages:

  • Intelligent, multi-attribute matching: Superlinked analyses jobs and candidates across skills, interests, location, seniority and company context, not just keywords. This richer view surfaces roles that truly fit each user and cuts role-mismatch complaints.
  • Real-time personalization: Every view, bookmark, application or dislike feeds straight back into the model, so recommendations update within seconds and always reflect a user’s current intent.
  • Higher engagement and conversions: Better matches keep users browsing longer, drive more bookmarks and lift applications by fifty percent.
  • Seamless integration and scalability: Superlinked’s API slotted into Climatebase’s stack with minimal effort and scales automatically as traffic and listings grow, adding no extra operational load.

In summary, switching to Superlinked gave Climatebase a more personalised, efficient job-matching engine. The AI-driven feed delivers higher view-to-application conversions, fewer irrelevant listings and stronger user engagement. Job seekers find roles that fit, employers reach qualified talent faster and Climatebase advances its mission to grow the climate workforce.

‍

Superlinked allows us to surface relevant recommendations to both sides of our marketplace - jobs to jobseekers and talent to employers - reducing friction and accelerating the growth of the climate workforce. Notably, SL's 'cold start' recommendations have been a major improvement for us, allowing us to show high quality recommendations to new users right away, before we even have any interaction data.

Yassine Hamdouni
Head of Engineering & Product
Posted by

Ben Gutkovich

COO & Co-founder

Share on social

Let’s launch vectors into production

Talk to Engineer
Subscribe to stay updated
You are agreeing to our Terms and Conditions by Subscribing.
Thank you!
Your submission has been received!
Oops! Something went wrong while submitting the form.
Copyright © 2025 Superlinked Inc. All rights reserved.