Ex-Spotify engineers raise $10M to kill the generic online storefront

Sidd Motwani, Ian Anderson, and Shivaditya Sinha spent years building the behavioral intelligence guts of Spotify’s recommendation engine, a system called Vector AI that predicts intent rather than just replaying your history. It now powers roughly 90% of the recommendations for Spotify’s 800 million users. The trio clearly saw a bigger game. Their new startup, Malachyte, just landed a $10 million seed round co-led by Bessemer Venture Partners and Gradient, with Harpoon Ventures joining in. The mission is straightforward and kind of damning for the state of e-commerce: most online stores treat you like a demographic stereotype or a walking receipt, especially if you’re not logged in. First-time visitors get a generic splash page, and loyal customers are stuck in a loop of past purchases.

Malachyte’s fix is a real-time, intent-aware platform built on what they’re calling “two-headed Vector AI.” The system starts profiling before you even click, reading the context of your visit immediately. Motwani, the CEO, laid out a concrete scenario to TechCrunch: a search for “heavy-duty boot” followed by a couple of clicks on steel-toed options is enough signal to dynamically push work pants and gloves up the page while burying the dress shoes. No account needed. No purchase history. Every subsequent hover, scroll, or add-to-cart sharpens the model’s understanding of both your general taste and your specific objective right that second.

What’s genuinely different here isn’t just the speed—it’s the refusal to batch-process intelligence. Motwani argues that most systems hoard behavioral signals and then aggregate them overnight into a stale segment. Malachyte reads them continuously, treating each action as a vote for the user’s current intent. The platform also weights contextual signals heavily, recognizing that a 11 p.m. phone browser clicking an email link is in a fundamentally different headspace than the same person on a work laptop at 10 a.m. That’s a level of nuance that static personalization engines simply ignore, and it’s exactly the kind of temporal thinking you’d expect from a team that mastered audio listening patterns.

The company tested this across over 20 enterprise clients in travel, grocery, and retail before zeroing in on e-commerce. It went live with Fun.com in fall 2025 and has been generally available to Shopify merchants through a native integration since June 2026, while larger players can tap in via API. Motwani’s ultimate play is to fuse merchandising and marketing onto a single, shared understanding of customer behavior—closing the absurd gap between what a store knows about you and how it actually treats you. Given that so much retail AI still acts like it’s 2015, a real-time challenger with Spotify-scale infrastructure cred is a genuine threat to the status quo.