At LinkedIn, enabling economic opportunity at scale is our job—we want to connect every member of the professional workforce in the world to opportunity. To help tackle this mammoth task, we use artificial intelligence (AI) to assist with everything from finding our members the right job openings to surfacing better candidates for our customers. With AI, we’re able to efficiently sort through the massive amount of data we have—job postings, people you may want to connect with, feed content, and more—and align recommendations with members’ interests.
We’ve been incorporating AI into our products and services for years. For a more detailed introduction of AI at LinkedIn, check out Deepak Agarwal’s blog post or the rest of our AI-related blogs.
From catching up on trending news to getting updates from your network, the LinkedIn Feed has become central to our member experience.
People You May Know
People You May Know (PYMK) allows members to grow their network by recommending new potential connections. It’s one of the most recognizable features at LinkedIn, was invented here, and is responsible for building more than 50% of LinkedIn’s professional graph.
Video: Deep Learning for Search and Recommendations
Deep learning has been widely successful in solving complex tasks such as image recognition (ImageNet), speech recognition, machine translation, etc. In the area of personalized recommender systems, deep learning has started showing promising advances in recent years. In this tutorial, we present ways to leverage deep learning towards improving recommender system.
Many of our Talent Solutions products are built around mathematical models that try answer a simple question: is this opportunity of interest to this member at this time?
Natural Language Processing
While much of our data is structured—graph nodes and edges, normalized fields in database records—a great deal of it is simply natural language text. Attaching structure and meaning to this text is essential to LinkedIn’s overall mission of connecting its members to opportunity.
QCon.ai 2019: PYMK: Fast Recommendations Over Massive Data
The principal challenge in developing a service like PYMK is dealing with the sheer scale of computation needed to make precise recommendations with a high recall. This talk dives into that and the evolution of PYMK to its current architecture.