recruiting · saas · workflow

JUDE: LinkedIn's LLM-based embedding platform for job recommendations

Deploying LLMs in production at LinkedIn's scale brought high computational costs, complex deployment pipelines, and continuous domain adaptation needs that the previous embedding platform Pensieve could not address efficiently.

How it works
Common implementation structure
How this type of workflow is generally built, generalized across documented cases — not tied to any one vendor's stack. Click any stage to read what happens there. Specific products that implement these stages appear in “Tools commonly seen” below.
Stage 1 · Entity change triggers pipeline
Separate Kafka and Brooklin streams representing the changelog for job postings, member profiles, and member resumes trigger embedding inference for their respective entities.
Tools used
PyTorchHugging FaceKafkaBrooklinSamzaVeniceHDFSModel CloudDeepSpeedLigerLoRAFlash attention 2FlyteSparkRayKubernetes
Outcome

JUDE embeddings ramped online replaced standardized features in job recommendation and search L2 ranking models, delivering +2.07% Qualified Applications, -5.13% Dismiss to Apply, and +1.91% Total Job Applications — described as the highest metric improvement from a single model change the team had observed that half year.

What failed first

The previous embedding platform Pensieve relied on imprecise smaller ML models, hard-to-maintain taxonomies, and rigid upstream pipelines using Lambda architecture, which caused time-travel data consistency issues and required monitoring and recovery of failed scheduled inference jobs.

Results
Time savedunder 300ms
Volume+2.07%
Cost replacedup to 3x
Source

https://www.linkedin.com/blog/engineering/ai/jude-llm-based-representation-learning-for-linkedin-job-recommendations

How we source this →

Grounding & classification
Source type: technical build writeup
40 fields verified against source quotes, 1 dropped as unverifiable.
enterprise searchpersonalizationrecommendation systemknowledge baseresumebuilder submittedmetric backednamed customerproduction runtime claimedtools describedworkflow describedsoftwareaccuracy improvementconversion increasecost reductiontechnical build writeuprecruitingdata sync enrichment