Former Google and Apple Researchers Launch a Startup to Build AI’s Missing Feedback Loop | WIREDSkip to main contentMenuSECURITYPOLITICSTHE BIG STORYBUSINESSSCIENCECULTUREREVIEWSMenuAccountAccountNewslettersSecurityPoliticsThe Big StoryBusinessScienceCultureReviewsChevronMoreExpandThe Big InterviewMagazineEventsWIRED InsiderWIRED ConsultingNewslettersPodcastsVideoLivestreamsMerchSearchSearchMaxwell ZeffBusinessMay 27, 2026 10:00 AMFormer Google and Apple Researchers Launch a Startup to Build AI’s Missing Feedback LoopTrajectory is betting the rapid iteration cycle that supercharged vibe-coding can help all kinds of companies build AI products that learn continuously.Trajectory founders, Ronak Malde (left), Michael Elabd(center), and Arjun Karanam (right).Photo-Illustration: WIRED Staff; Courtesy of Trajectory AICommentLoaderSave StorySave this storyCommentLoaderSave StorySave this storyA group of AI researchers who previously worked at Google DeepMind, Apple, OpenAI, and Meta Superintelligence Labs announced on Wednesday they’re launching a new startup called Trajectory, which aims to help companies regularly improve their AI products by training on real-world user interactions.Trajectory wants to build a platform for AI that can learn continuously, a capability that researchers have long held up as a major barrier to further AI progress. OpenAI, Google, and Anthropic have found success training increasingly capable versions of AI models, especially for domains such as coding, math, and science. However, these systems stop getting smarter after their training is done. While there have been some recent breakthroughs in continual learning, tech companies have generally struggled to make AI products that learn from their errors in real time. In December 2025 at NeurIPS, one of the largest annual AI research conferences, Turing award winner Richard Sutton argued that continual learning is essential for building superintelligent agents.Trajectory has raised a $15 million seed round at a $115 million post-money valuation, led by the venture capital firm Conviction, with participation from Bessemer Venture Partners, Radical VC, and BoxGroup. Individual investors also participated in the round, including Google DeepMind’s chief scientist, Jeff Dean, as well as the so-called “godmother of AI,” Stanford professor and World Labs CEO Fei-Fei Li.Trajectory’s CEO and cofounder Ronak Malde was previously an AI researcher at Windsurf, and he later became one of only a handful of employees who went to work at Google DeepMind when it hired the coding startup’s top talent in a $2.4 billion deal last year. The other cofounders of Trajectory include Arjun Karanam, a former AI researcher at Apple who worked on the Vision Pro, and Michael Elabd, who previously worked in Google DeepMind’s robotics division.Malde tells WIRED that some leading AI coding products, such as Cursor, are already doing an early version of continual learning—using real data about how people interact with their products to do post-training and regularly ship model improvements. He argues this is a core reason why AI coding products have taken off so rapidly, and is part of the reason why major AI labs have rushed to develop vibe coding applications of their own. With Trajectory, Malde and his team of 11 researchers and engineers hope to apply a similar technique for improving AI-powered tools outside the coding space.“Even the most powerful AI today is still static. The AI model that you used yesterday is going to make the same mistakes today,” says Malde. “A couple companies are starting to get to that world of continual learning. What we are doing is building the platform for every single company to get to continual learning.”The challenge with applying this logic to other domains is that coding is easily verifiable—code either runs or it doesn’t—but some industries have looser definitions of success. Karanam says part of what Trajectory’s platform offers is helping optimize an AI model to a business's specific needs.Rather than starting from an off-the-shelf model from OpenAI or Anthropic, Trajectory has customers begin with an open-source model that has been post-trained for a specific AI product the company has in mind. For Decagon, a customer that builds AI customer support agents, Trajectory logs when its AI falls short—say, a customer trying to make a return gets their query bounced to a human—and uses those instances to post-train a new model as often as every week. Trajectory claims these post-trained models beat the frontier labs’ models on narrow tasks that matter most for a company’s product.Corporate executives are eager to use AI for many different kinds of tasks, but to do that today they often need to hire teams of “forward deployed engineers,” or consultants and technical employees embedded inside a company who help build out AI products. Companies like OpenAI, Anthropic, and Palantir have rushed to fill that need. Elabd says Trajectory’s goal is to build a product that can improve on its own so that companies don’t need in-house engineers to continuously troubleshoot their AI stack. The startup says it has customers in a variety of fields already, including the enterprise sales startup Clay and the legal AI startup Harvey. While it currently works primarily with AI-native companies, Trajectory eventually plans to market its platform to the Fortune 500.Critics could argue that Trajectory has not yet built true continual learning, at least not in the traditional sense. After all, the startup’s models only update once a week at this time, and they remain static between upgrades.Elabd argues that Trajectory is just getting started. He claims the AI industry is moving towards a new paradigm where AI learns from experience—much like what’s already happening in the AI coding space. Elabd says Trajectory’s eventual goal is to build a platform that can update a company’s AI model every single day, or perhaps even more frequently.“Every day may not be enough. It could be every hour, it could be every interaction,” says Elabd. “Maybe every company doesn’t need just one AI, you could train an AI to learn for every person at every company.”This is an edition of Maxwell Zeff’s Model Behavior newsletter. Read previous newsletters here.CommentsBack to topTriangleYou Might Also LikeHow to find us: Add WIRED.com to your preferred sources in GoogleThese women are trying to optimize their vaginasBig Story: AI gig work is the new waiting tables—and it's soul-crushingThis summer, the American water crisis becomes realEvent: How to adapt, compete, and win in the next era of businessMaxwell Zeff is a senior writer at WIRED covering the business of artificial intelligence. He was previously a senior reporter with TechCrunch, where he broke news on startups and leaders driving the AI boom. Before that, Zeff covered AI policy and content moderation for Gizmodo and wrote some of Bloomberg’s ... Read MoreSenior WriterTopicsModel Behaviorartificial intelligenceStartupsgenerative AIVenture CapitalRead MoreNick Bostrom Has a Plan for Humanity’s ‘Big Retirement’The philosopher thinks humans should pursue advanced AI and the promise of a “solved world.”Steven LevyElon Musk Seemingly Admits xAI Has Used OpenAI’s Models to Train Its OwnWhile answering questions under oath, Musk argued it’s standard practice for AI labs to use their competitors’ models.Maxwell ZeffOpenAI Really Wants Codex to Shut Up About Goblins“Never talk about goblins, gremlins, raccoons, trolls, ogres, pigeons, or other animals or creatures unless it is absolutely and unambiguously relevant,” reads OpenAI’s coding agent instructions.Will KnightTop Surfshark Promo Codes for June 2026Save up to 87% with a Surfshark coupon code, 3 months of VPN free today, and more from WIRED.Scott GilbertsonLiterary Prizewinners Are Facing AI Allegations. It Feels Like the New NormalThree of five regional winners of the prestigious Commonwealth Short Story Prize are suspected of relying on chatbots. They’re certainly not alone.Miles KleeI Am Begging AI Companies to Stop Naming Features After Human ProcessesAnthropic announced “dreaming” for AI agents to sort through “memories” at its developer conference. Can we not?Reece RogersSubmit Your Questions: AI Is Changing Your Job—Now What?Pose your questions ahead of our May 27 livestream AMA, where a panel of WIRED experts will discuss how AI is transforming work.Reece RogersUsing AI for Just 10 Minutes Might Make You Lazy and Dumb, Study ShowsNew research suggests that reliance on AI assistants can have a negative impact on people’s ability to think and problem solve.Will KnightChatGPT Has ‘Goblin’ Mania in the US. In China It Will ‘Catch You Steadily’OpenAI’s chatbot has some weird linguistic tics in Chinese that are driving users crazy.Zeyi YangThe New Wild West of AI Kids’ ToysThese cuddly, connected companions could disrupt everything from make-believe to bedtime stories. No wonder some lawmakers want them banned.Sophie ChararaTop B&H Photo Promo Codes and Deals This JuneEnjoy top deals on cameras, computers, and tech essentials at B&H Photo.Scott GilbertsonLitter-Robot Promo Codes: $150 Off Bundles, $50 Off the Starter Bundle, and MoreGet the latest Litter-Robot Discounts on Litter-Robot self-cleaning litter boxes, accessories, and more.Molly HigginsWIRED is obsessed with what comes next. Through rigorous investigations and game-changing reporting, we tell stories that don’t just reflect the moment—they help create it. When you look back in 10, 20, even 50 years, WIRED will be the publication that led the story of the present, mapped the people, products, and ideas defining it, and explained how those forces forged the future. WIRED: For Future Reference.More From WIREDSubscribeNewslettersLivestreamsTravelFAQWIRED StaffWIRED EducationEditorial StandardsArchiveRSSSite MapAccessibility HelpReviews and GuidesReviewsBuying GuidesStreaming GuidesWearablesCouponsGift GuidesAdvertiseContact UsManage AccountJobsPress CenterCondé Nast StoreUser AgreementPrivacy PolicyYour California Privacy Rights© 2026 Condé Nast. All rights reserved. WIRED may earn a portion of sales from products that are purchased through our site as part of our Affiliate Partnerships with retailers. The material on this site may not be reproduced, distributed, transmitted, cached or otherwise used, except with the prior written permission of Condé Nast. Ad ChoicesSelect international siteUnited StatesLargeChevronItaliaJapónCzech Republic & SlovakiaFacebookXPinterestYouTubeInstagramTiktok

A group of AI researchers from former positions at Google DeepMind, Apple, OpenAI, and Meta Superintelligence Labs has launched a new startup named Trajectory, which aims to establish a platform enabling AI systems to continuously improve by training on real-world user interactions. This initiative addresses the long-standing barrier in AI progress concerning the capacity for continual learning, a concept deemed essential by researchers like Richard Sutton for building superintelligent agents. While major AI labs have achieved success by training increasingly capable models in domains like coding and mathematics, these systems typically become static after initial training, leaving a gap in the ability for AI products to learn from operational errors in real time. Trajectory proposes a solution by betting on rapid iteration cycles, similar to those seen in vibe coding, to facilitate continuous learning across various industries.

The founders of Trajectory include Ronak Malde, Arjun Karanam, and Michael Elabd, who bring experience from these leading AI organizations. Malde notes that current AI models are inherently static, meaning the model from yesterday will make the same mistakes today; Trajectory seeks to build the infrastructure for every company to enter this realm of continual learning. While some existing AI coding products, such as Cursor, have implemented early versions of continual learning by using real product interaction data for post-training model improvements, Trajectory intends to extend this methodology beyond coding. Karanam suggests that a key benefit of their platform is optimizing AI models to specific business requirements, moving beyond off-the-shelf models from entities like OpenAI or Anthropic.

The operational mechanism involves starting with open-source models that have already been post-trained for a specific application. For instance, a customer like Decagon, which develops AI customer support agents, uses Trajectory to log instances where the AI fails, such as when a customer query is mistakenly bounced to a human, and subsequently uses these failures to post-train a new model on a weekly basis. Trajectory claims these custom post-trained models can outperform frontier lab models on narrow tasks critical to a company’s product.

The startup’s overarching goal is to reduce the dependency on hiring in-house engineers to continuously troubleshoot AI stacks, aspiring to build a product that can self-improve. They have already engaged customers in various sectors, including enterprise sales startups like Clay and legal AI firms like Harvey, with plans to eventually market the platform to the Fortune 500. Although the current customer base is largely AI-native companies, Trajectory plans broader market expansion.

Despite this ambition, critics raise concerns that Trajectory has not yet achieved true continual learning in the traditional sense, pointing out that the models currently only update weekly and remain static between upgrades. However, the founders argue that the industry is shifting towards an experience-based learning paradigm, and Trajectory is building the necessary platform for this evolution. They assert that the AI industry is moving toward a model where learning from experience is standard, and their ultimate vision is to create a platform capable of updating a company’s AI model daily, or even more frequently, eventually allowing an AI to learn from every interaction across every company.