Semantic Search For Ai Agents At Scale Retrieval And Ranking For Linkedins Hiring Assistant

Semantic retrieval for an agent needs a scoring rubric before it needs a model: LinkedIn defined "qualified", built an AI judge to that standard, then trained embeddings on those judgments to rank 1.3B profiles in real time.

• eval — A repeatable test for AI quality — a set of tasks plus scoring — used the way software teams use test suites, because model output is too variable to judge by eyeballing.
• embedding — A list of numbers representing a piece of text's meaning, so that similar meanings end up numerically close and can be searched.
• AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
• LLM-as-judge — Using one model to score another's output against a rubric, so quality can be measured at a scale human grading cannot reach.

• articleInside Linkedins Data Context Layer Curating Trustworthy Context For Humans And Agents
• articlePensieve
• articleEvaluating Success Of Consumer Generative Ai Products
• articleBuilding A Large Scale Recommendation System People You May Know

• articleThe Tech Behind The First Agent From Linkedin Hiring AssistantLinkedIn Engineering
• articleBuilding The Next Generation Of Job Search At LinkedinLinkedIn Engineering
• articleHow We Engineered Linkedins Hiring AssistantLinkedIn Engineering

Checking sign-in…

Loading comments…