Jev Gains Attention for Search and Recommendation After Perfectly Tests It on ECCV Researcher Matching

A recruiting-search founder and former TikTok machine-learning engineer says TypeSafe AI's Jev, combined with Perfectly, was used to analyze 500 ECCV 2026 researchers through their papers while reaching the performance of one Claude model on the task, putting a new spotlight on decision-only AI for search and recommendation systems. The claim is a narrow application example rather than a public benchmark, but it lands days after TypeSafe introduced Jev as an early-access "System One" model built to return typed probabilistic decisions instead of generated text.