
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 ...
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...
01/06/2026
January 6 2026, 05:30 (PST) Dolby Sets the New Standard for Premium Entertainment at CES 2026
Throughout the week, Dolby brings to life the latest innovatio...
02/05/2026
Dalet, a leading technology and service provider for media-rich organizations, t...
01/05/2026
January 5 2026, 18:30 (PST) NBCUniversal's Peacock to Be First Streamer to ...
23/04/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
23/04/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
23/04/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
23/04/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
23/04/2026
Partnership between ARRI and SmallHD brings new Hi-5 license
Configurable monitor overlays adapt to individual working styles
Supported by SmallHD monitors ru...
23/04/2026
Lighting Master Cronenweth ASC brings a unique look to each grid world with the help of Astera
Jeff Cronenweth on the set of Disney's TRON: ARES. Photo by...
23/04/2026
DP Chloe Smolkin ( The Late Show, Kidz Bop ) joins director Danielle Beckmann and writer/actor Raji Ahsan behind the camera for the heartfelt short comedy Dr...
23/04/2026
Apply now to join the 2026 Producer Delegation to TIFF: The Market 23 April 2026
Screen Australia, in partnership with Ontario Creates, has opened application...
22/04/2026
Solid State Logic is advancing its System T platform with a stronger focus on IP...
22/04/2026
From immersive audio to live streaming, Dolby Laboratories is focused on the fut...
22/04/2026
Shallow depth-of-field cameras have taken the industry by storm. Its debut a han...
22/04/2026
Riedel Communications (Booth C4908) announced that Eastern Kentucky University (...
22/04/2026
The NAB Show is in full swing, and the SVG and SVG Europe editorial teams are chasing down the hottest stories from all over the Las Vegas Convention Center. He...
22/04/2026
Blackmagic Design has announced the URSA Cine 12K LF 100G, a new model in the URSA Cine family adding 100G Ethernet for SMPTE 2110 live production output up to ...
22/04/2026
Celebrating its 40th anniversary, NEP is leaning into hybrid production with the...
22/04/2026
NEP VP, Platform Dan Murphy sits down at the 2026 NAB Show to unpack what NEP P...
22/04/2026
Spotify and the New York Liberty are teaming up to give music and basketball fan...
22/04/2026
New 20-minute documentary explores iconic design The Focusrite Room in Mesa, Arizona, where John Aquilino hosts the Studio Console 005.
In 2025, Focusrite co...
22/04/2026
Offers compact wireless solution for pedalboards
Taiwanese audio brand Cloudvocal have announced the availability of a new pedalboard-friendly wireless syst...
22/04/2026
Latest hybrid sampling/synthesis instrument arrives
Arturia's Augmented series offerings rely on a mixture of sampling and synthesis, allowing users to ...
22/04/2026
Combines three distinct analogue EQ emulations
The latest addition to Acustica Audio's ever-expanding collection of analogue-emulation plug-ins combines...
22/04/2026
Final instalment in vintage-inspired instrument series
Analog Empire: Bass & Lead marks the final instalment in Melda Production's vintage hardware-insp...
22/04/2026
Fuzz pedal joins all-analogue Series A line
Given that Strymons reputation was built on unapologetically digital pedals, it was a little surprising to see t...
22/04/2026
SBS names shortlisted brands for 2026 SBS Media Sustainability Challenge
22 April, 2026
Media releases
National broadcaster also releases its second annual...
22/04/2026
Why Low Band Electronic Warfare Matters...
22/04/2026
The nation unites around football team's World Cup dream
Warsaw, Poland, 20.04.26: Nielsen, a global leader in audience measurement, data, and media intell...
22/04/2026
Warsaw, Poland, 22.04.26: Nielsen, a global leader in audience measurement, data...
22/04/2026
New market intelligence offering gives businesses a clearer view of local consum...
22/04/2026
Glookast Unveils New UX, YouTube and Social Media Connectors, Premiere Panel, Ci...
22/04/2026
Lightcraft Technology to Preview Spark Story at NAB 2026 with Interactive Previs...
22/04/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
22/04/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
22/04/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
22/04/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
22/04/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
22/04/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
22/04/2026
22 Apr 2026
VEON's Banglalink to Bring Starlink Mobile to Customers in Bangladesh Bangladesh becomes the third market where VEON and Starlink Mobile partne...
22/04/2026
U have unveiled exclusive first-look images for their six-part police thriller Hit Point, starring Nick Blood (Day of the Jackal) and BAFTA nominee Saffron Hock...
22/04/2026
What can I watch on UKTV and stream on U this week?
This week on UKTV and the free streaming service U, viewers can watch a range of new and returning programm...
22/04/2026
Wednesday 22 April 2026
Sky announces fifth year of WNT Fund with 30,000 bursa...
22/04/2026
Back to All News
This Earth Day, Discover the Sustainable Productions Behind Our Films and Series
Emma Stewart, Ph.D.
Netflix Sustainability Officer
Enterta...
22/04/2026
The move from Retail Media to Commerce Media is about broadening the scope of th...
22/04/2026
April 22 2026, 07:00 (PDT) Dolby and BMW Bring Dolby Atmos to the BMW 7 Series,...
22/04/2026
RT Documentary On One 7-part series breaks US market for first time
RT Programme Sales has announced its first deal with a US distribution partner for its 7-...