
Pinterest has engineered a way to serve its photo-sharing community more of the images they love.
The social-image service, with more than 400 million monthly active users, has trained bigger recommender models for improved accuracy at predicting people's interests.
Pinterest handles hundreds of millions of user requests an hour on any given day. And it must also narrow down relevant images from roughly 300 billion images on the site to roughly 50 for each person.
The last step - ranking the most relevant and engaging content for everyone using Pinterest - required a leap in acceleration to run heftier models, with minimal latency, for better predictions.
Pinterest has improved the accuracy of its recommender models powering people's home feeds and other areas, increasing engagement by as much as 16%.
The leap was enabled by switching from CPUs to NVIDIA GPUs, which could easily be applied next to other areas, including advertising images, according to Pinterest.
Normally we would be happy with a 2% increase, and 16% is just a beginning for home feeds. We see additional gains - it opens a lot of doors for opportunities, said Pong Eksombatchai, a software engineer at Pinterest.
Transformer models capable of better predictions are shaking up industries from retail to entertainment and advertising. But their leaps in performance gains of the past few years have come with a need to serve models that are some 100x bigger as their number of model parameters and computations skyrockets.
Huge Inference Gains, Same Infrastructure Cost Like many, Pinterest engineers wanted to tap into state-of-the-art recommender models to increase engagement. But serving these massive models on CPUs presented a 100x increase in cost and latency. That wasn't going to maintain its magical user experience - fresh and more appealing images - occurring within a fraction of a second.
If that latency happened, then obviously our users wouldn't like that very much because they would have to wait forever, said Eksombatchai. We are pretty close to the limit of what we can do on CPU basically.
The challenge was to serve these hundredfold larger recommender models within the same cost and latency constraints.
Working with NVIDIA, Pinterest engineers began architectural changes to optimize their inference pipeline and recommender models to enable the transition from CPU to GPU cloud instances. The technology transition began late last year and required major changes to how the company manages workloads. The result is a 100x gain in inference efficiency on the same IT budget, meeting their goals.
We are starting to use really, really big models now. And that is where the GPU comes in - to help make these models possible, Eksombatchai said.
Tapping Into cuCollections Switching from CPUs to GPUs required rethinking its inference systems architecture. Among other issues, engineers had to change how they send workloads to their inference servers. Fortunately, there are tools to assist in making the transition easier.
The Pinterest inference server built for CPUs had to be altered because it was set up to send smaller batch sizes to its servers. GPUs can handle much larger workloads, so it's necessary to set up larger batch requests to increase efficiency.
One area where this comes into play is with its embedding table lookup module. Embedding tables are used to track interactions between various context-specific features and interests of user profiles. They can track where you navigate, and what people Pin on Pinterest, share or numerous other actions, helping refine predictions on what users might like to click on next.
They are used to incrementally learn user preference based on context in order to make better content recommendations to those using Pinterest. Its embedding table lookup module required two computation steps repeated hundreds of times because of the number of features tracked.
Pinterest engineers greatly reduced this number of operations using a GPU-accelerated concurrent hash table from NVIDIA cuCollections. And they set up a custom consolidated embedding lookup module so they could merge requests into a single lookup. Better results were seen immediately.
Using cuCollections helped us to remove bottlenecks, said Eksombatchai.
Enlisting CUDA Graphs Pinterest relied on CUDA Graphs to eliminate what was remaining of the small batch operations, further optimizing its inference models.
CUDA Graphs helps reduce the CPU interactions when launching on GPUs. They're designed to enable workloads to be defined as graphs rather than single operations. They provide a mechanism to launch multiple GPU operations through a single CPU operation, reducing CPU overheads.
Pinterest enlisted CUDA Graphs to represent the model inference process as a static graph of operation instead of as those individually scheduled. This enabled the computation to be handled as a single unit without any kernel launching overhead.
The company now supports CUDA Graph as a new backend of its model server. When a model is first loaded, the model server runs the model inference once to build the graph instance. This graph can then be run repeatedly in inference to show content on its app or site.
Implementing CUDA Graphs helped Pinterest to significantly reduce inference latency of its recommender models, according to its engineers.
GPUs have enabled Pinterest to do something that was impossible with CPUs on the same budget, and by doing this they can make changes that have a direct impact on various business metrics.
Learn about Pinterest's GPU-driven inference and optimizations at its GTC session, Serving 100x Bigger Recommender Models, and in the Pinterest Engineering blog.
Register for GTC, running Sept. 19-22, for free to attend sessions with NVIDIA and dozens of industry leaders.
Most recent headlines
05/01/2027
Worlds first 802.15.4ab-UWB chip verified by Calterah and Rohde & Schwarz to be ...
06/09/2026
June 9 2026, 23:00 (PDT) Dolby and MagentaTV Bring Fans Closer to the FIFA Worl...
04/08/2026
Dalet, a leading technology and service provider for media-rich organizations, t...
04/07/2026
April 7 2026, 19:00 (PDT) Detective Conan: Fallen Angel of the Highway Opens in...
17/06/2026
EVS has announced it has received the EcoVadis Gold Medal for sustainability performance, ranking among the top 5% of companies globally in the Technology/Mid-S...
17/06/2026
Chyron has released Weather 2.4, an update to its weather suite for broadcasters and meteorologists. The release focuses on enhancements to the DataFlow module,...
17/06/2026
VSiN, The Sports Betting Network, has announced the launch of Best Bets TV, a free ad-supported streaming TV (FAST) channel. The 24/7 channel is currently avail...
17/06/2026
DAZN's Team Whistle and Snap Inc. have announced a creator program centered ...
17/06/2026
LiveU is providing video transmission technology for broadcasters, production companies, and public safety agencies across North America's busy Summer of So...
17/06/2026
The National Academy of Television Arts and Sciences (NATAS) has announced that Laurens Grant and Jacob Ullman have joined its Board of Directors. Chief of Staf...
17/06/2026
Akta, the AI-First SaaS video platform for modern broadcast and streaming operations, today announced that its video platform is now generally available on Orac...
17/06/2026
This recent graduate from Houston found inspiration in technical directing and now eyes a future career in sports production...
17/06/2026
Audio-Technica (booth C7959) arrives at InfoComm 2026 in Las Vegas with a slate ...
17/06/2026
SNS has published a guide addressing growing demand for AI-powered video indexing, transcription, facial recognition, and searchable metadata across media libra...
17/06/2026
Providius has announced Providius Direct, a workflow for investigating network i...
17/06/2026
NEP Group has announced the commercial availability of NEP Platform, a software ...
17/06/2026
A new white paper examining Secure Reliable Transport (SRT) and Reliable Interne...
17/06/2026
New research from subscription bundling platform Bango finds that younger sports fans are increasingly consuming sport through highlights, clips, and social med...
17/06/2026
SMPTE has announced that its complete Standards catalog is now freely available to the global media technology community, including all published SMPTE Standard...
17/06/2026
Harmonic has completed the sale of its Video Business to MediaKind for $145 mill...
17/06/2026
Omaha Productions will produce the 2026 World Series of Poker (WSOP) in Las Vega...
17/06/2026
New features, changes & bug fixes
SoundBridge have just released another update for their remote collaboration-focused DAW - reviewed here in SOS March 2026...
17/06/2026
Valve-based front end for digital & modelling rigs
The latest addition to Fryette's product range delivers a packed-down, pedalboard-friendly version of...
17/06/2026
GearExpo UK - 27 June 2026
Sound On Sound are proud to announce GearExpo UK, a major new recording and music technology exhibition in London! This is the bi...
17/06/2026
SAM monitoring line-up gains Dante and AES67 support
The latest expansion of Genelec's UNIO monitoring ecosystem introduces a new device that provides D...
17/06/2026
The R&S PR300 portable receiver from Rohde & Schwarz sets new standards in spect...
17/06/2026
Elt Group and Rohde & Schwarz sign a cooperation agreement to explore commercial...
17/06/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
17/06/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
17/06/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
17/06/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
17/06/2026
The Immersive Supervisor Emerges as Hollywood's Next Production Role
Brie Clayton June 17, 2026
0 Comments
Above image: On set live immersive revi...
17/06/2026
Vertical Musical Playback Shot with Blackmagic PYXIS 6K
Brie Clayton June 17, 2026
0 Comments
Large format sensor and DaVinci Resolve workflow used fo...
17/06/2026
DAZ 3D Launches New Game-Ready Character Assets Built for Modern Engines and Pro...
17/06/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
17/06/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
17/06/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
17/06/2026
Montr al, Quebec, June 11, 2026 Kiloview, a leading provider of AV-over-IP and NDI -based video transmission solutions, today announced a distribution partner...
17/06/2026
Changsha, China, June 15, 2026 Kiloview officially announced the launch of U4 IP Video Dock, a compact IP video decoder and output dock designed to bring prof...
17/06/2026
Save 50% on All Ivory II Pianos and Collections
Were halfway through the year, and we're cutting prices in half for the Ivory II Summer Sale!
For a limite...
17/06/2026
Wednesday 17 June 2026
Two in three fans will connect to venue WiFi this World ...
17/06/2026
Visibility builds credibility - the tools you use every day, now visible on your LinkedIn profile Published on Jun 17, 2026 Categories: Company News, Product ...
17/06/2026
Transaction Positions Harmonic as a Pure-Play Broadband Company SAN JOSE, Calif. - June 17, 2026 - Harmonic Inc. (NASDAQ: HLIT), the worldwide leader in virtual...
17/06/2026
FOX Advertising To Launch Industry's First End-to-End Agentic Advertising Pl...
17/06/2026
How SGN is future-proofing critical national infrastructure with Arqiva Managed Connectivity.
When disruption becomes the norm As Storm Eunice tore across the ...
17/06/2026
The RT Concert Orchestra will bring the timeless music of The Beach Boys to aud...
17/06/2026
Calibr-Skaggs awarded $5.1M by NIH to develop long-acting hepatitis B virus therapy A new program aims to replace a daily HBV drug with once-monthly or even qua...
16/06/2026
Thomson's highly regarded expert-led online learning courses are now easier to access on the go via our new App.
Available now on Google Play Store, the J...
16/06/2026
Neumann.Berlin has released firmware version 2.0 for the MT 48 audio interface, adding plugin compatibility, expanded Dante networking options, broadcast encode...
16/06/2026
TVNewsCheck has announced that nominations are now open for its 2027 Women in Technology Awards, to be presented at NAB Show 2027 on Tuesday, April 6 in the Med...