Machine Learning Engineer I

Pinterest increased ad relevance by 300%. May 19, 2026. Pinterest has announced a major update to the way its advertising system works, helping businesses show more relevant ads to users at the right time. The platform has improved its ad-serving model so it can now combine a user's previous online activity with what they are currently searching for on Pinterest. This means users are more likely to see products and promotions that genuinely match their interests, while businesses could benefit from stronger campaign performance and better engagement. In a recent update shared on the Pinterest Engineering blog, the company explained that its system already used external conversion history to help predict future customer behaviour. In simple terms, Pinterest could look at actions users had taken previously, such as purchases or website visits, to help decide which ads they may be interested in. While this was useful, it only looked at past behaviour and didn't consider what users were actively searching for in the moment. What's the improvement? To improve this, they have now introduced a new system that also tracks real-time activity on the platform. This allows Pinterest to better understand what users are interested in right now, not just what they have engaged with in the past. By combining these two sets of information, Pinterest can deliver adverts that feel more personalised and relevant. The update focuses on three key areas: a new model architecture, a new training method, and a hybrid serving flow. While the technical details are complex, the overall goal is straightforward to improve the quality and relevance of the ads users see. One of the biggest improvements comes from the way the platform now analyses "Related Pins". These are the suggested pins users see while browsing content. The updated system can examine the pin a user is currently viewing and identify the topics and interests connected to it. Pinterest then uses this information, alongside previous browsing and purchase behaviour, to predict which promoted content is most likely to appeal to that user. For example, if someone is browsing home office inspiration and has recently visited websites selling office furniture, Pinterest can now connect those signals together far more effectively. As a result, that user may be shown adverts for desks, office chairs, or workspace accessories that closely match their current interests. Pinterest says this improved system has already shown impressive results during testing. The company reported that ad relevance improved between three and ten times compared to previous models. It also revealed that the "median relevance" of ads increased by around 275% to 300%. In addition, Pinterest saw a rise in the number of suitable ads being delivered to users. The platform stated that it was able to retrieve twice as many relevant ad candidates for impressions, meaning users were shown a wider range of content that matched their interests. Although much of the announcement is highly technical, the key takeaway for businesses is clear: Pinterest is becoming smarter at matching products and services with the right audiences. This could lead to stronger engagement, improved click-through rates, and potentially better return on investment for advertisers using Pinterest campaigns. The update also highlights Pinterest's continued investment in artificial intelligence and machine learning technology. These tools are helping the platform improve content discovery, personalise user experiences, and create more accurate advertising opportunities. Pinterest is already seeing strong growth across the platform. The company recently announced it now has 631 million monthly active users worldwide. This represents an increase of 60 million users compared to the previous year, showing that more people are turning to Pinterest for inspiration, product discovery, and shopping ideas. For businesses, this growth presents a valuable opportunity. As Pinterest continues to improve its advertising tools and user experience, brands have a greater chance of reaching highly engaged audiences who are actively searching for products, ideas, and inspiration. The latest advertising improvements could be especially useful for retailers, lifestyle brands, service providers, and e-commerce businesses looking to connect with customers at different stages of the buying journey. By using both historical data and real-time behaviour, Pinterest is aiming to make advertising feel less intrusive and more helpful for users. As digital marketing becomes increasingly competitive, platforms that can deliver highly relevant and personalised content are likely to see stronger results for advertisers. Pinterest's latest update is another sign that AI-powered targeting and smarter audience insights are becoming a key part of modern marketing strategies.

Snapchat owner cuts 1,000 jobs as says AI will reduce repetitive work. 1 hour ago Kali HaysTechnology reporter Snap is the latest tech company to cut jobs, citing the growing ability of artificial intelligence (AI) tools. The Snapchat owner has axed about 1,000 employees - 16% of staff - and withdrawn hundreds of open job roles, according to a financial disclosure. Evan Spiegel, Snap's co-founder and CEO, told employees the company was in "a crucible moment" and the goal of the cuts was to reduce yearly costs by $500m (£368m). Spiegel said that workers who remain at the company will be using AI tools to "reduce repetitive work and increase velocity," as "small squads" of employees have been doing in recent months. "Change of this magnitude and at this speed is never easy and it will not be seamless," Spiegel added in his memo. This cut at Snap marks at least the third larger scale layoff at the company since 2022, when it conducted its first major layoff which impacted 20% of its staff at the time. Spiegel's comments on Wednesday are the first time he has pointed to AI as an explanation for staffing decisions. This year, activist investor Irenic Capital Management took a stake in Snap, saying in a public letter to Spiegel that it was "strange" the company remained unprofitable after 15 years in business and with hundreds of millions of monthly users. Irenic noted that an investor who put $1 into Snap when it went public in 2017 would be left with a stake worth only 23 cents today. An activist investor is a person or investment firm that buys shares in a company it believes is underperforming and then applies pressure for management and business changes. Snap now requires "a new way of working that is faster and more efficient, while pivoting towards profitable growth", Spiegel wrote in his memo to staff. His explanation echoed that of other tech executives this year who have begun citing the increased capability of AI technology, mainly tools that help software engineers do coding work, as they order their own mass job cuts. Already this year Amazon, Meta, Block, Pinterest and Atlassian, among others, have collectively laid off several thousand workers. Company executives have either pointed to increased use of AI tools meaning they need fewer workers, or noted that plans to spend hundreds of billions of dollars a year on investing in AI requires cost-cutting in other parts of the company. Jack Dorsey, the CEO of Block and the former leader of Twitter, said at the end of February that the rise of AI tools for tech workers "fundamentally changes what it means to build and run a company". He also said people in the industry should expect further job cuts at "the majority of companies" over the next year.