Former Spotify Engineers Launch Malachyte to Reinvent E-Commerce Personalization
Three engineers who spent years building the behavioral intelligence infrastructure behind Spotify's recommendation engine have now turned their attention to online retail. Sidd Motwani, Ian Anderson, and Shivaditya Sinha — the team behind a system called Vector AI that powers approximately 90% of Spotify's recommendations to its 800 million users — have founded a new startup called Malachyte. The company announced a $10 million seed funding round to scale distribution and expand its product and commercial teams. The round was co-led by Bessemer Venture Partners and Gradient, with participation from Harpoon Ventures.
The Problem: E-Commerce Personalization Is Stuck in the Past
Malachyte was founded on a specific diagnosis of how most online stores currently handle personalization. According to the founders, the prevailing approach relies on three inputs: historical purchases, demographic segmentation, and logged-in customer profiles. The consequence is predictable: first-time visitors are served generic storefronts, while returning customers receive recommendations based on what they bought before rather than what they are looking for right now.
This model, in the view of Malachyte's CEO Sidd Motwani, leaves significant value on the table. Retailers already collect a continuous stream of behavioral signals — hovers, clicks, scrolls, search refinements, and add-to-cart actions — but most systems either ignore these signals in the moment or only aggregate them overnight into broad customer segments. The result is a missed opportunity to act on the most current and specific expression of a shopper's intent.
The Technology: Two-Headed Vector AI
Malachyte's platform is built around what it describes as "two-headed Vector AI." The system is designed to simultaneously predict what product a shopper wants next and learn their general taste preferences, updating both models continuously in real time based on in-session behavior.
A key design principle is that the system does not require a user account or purchase history to begin operating. According to Motwani, the system starts forming a behavioral profile before the first click, using contextual signals available the moment a page loads.
Motwani illustrated the approach with a concrete example: "A search for 'heavy-duty boot' followed by two clicks on steel-toed boots is enough to move work pants and gloves up the page and push dress shoes down, with no account or history required. Every additional action sharpens the profile, so the experience gets more relevant the longer someone stays, and again on their next visit."
The platform also takes contextual signals into account that most systems ignore entirely. As Motwani noted, a phone visitor arriving at 11 p.m. via an email link is in a fundamentally different state of mind than the same person browsing on a laptop mid-morning — yet most e-commerce systems treat these two scenarios identically.
From Development to Live Deployment
Malachyte has been developing and testing its technology since 2024. Prior to focusing exclusively on e-commerce, the company worked with more than 20 enterprise customers across the travel, grocery, and retail sectors. The platform went live in fall 2025 with Fun.com as its first production deployment. Since June 2026, the platform has been generally available to Shopify merchants through a native integration. Larger retailers can also access the technology through a direct API integration.
Why This Matters for E-Commerce Operators
For shop operators, e-commerce managers, and developers working in the Shopware or broader Shopify ecosystem, Malachyte represents a notable shift in the architecture of recommendation systems. The conventional approach to personalization — building profiles from historical data and updating them in batch processes — is increasingly being challenged by real-time, session-aware systems that operate without requiring user authentication.
This has practical implications across several dimensions:
- Anonymous visitor monetization: Because the system builds a behavioral profile from the first page load without requiring a login, merchants can deliver relevant experiences to the significant share of visitors who never create an account.
- Reduced dependence on cookies and historical data: In an environment of increasing privacy regulation and cookie deprecation, intent-based real-time signals offer an alternative signal source that does not depend on persistent identifiers.
- Merchandising and marketing alignment: Motwani has stated that the longer-term opportunity lies in unifying merchandising and marketing decisions around a shared, real-time understanding of customer behavior — a goal that has long been technically difficult to achieve.
Practical Takeaways for Merchants and Developers
If you operate a Shopify store or are evaluating personalization infrastructure for a mid-to-large e-commerce operation, there are several considerations worth drawing from Malachyte's approach:
- Audit your current signal utilization: Identify whether your existing analytics and personalization stack acts on behavioral signals in real time or batches them for overnight processing. The gap between these two approaches is where intent-based systems claim their advantage.
- Evaluate anonymous visitor share: If a large proportion of your traffic never logs in, personalization systems that require a profile may be delivering generic experiences to the majority of your audience.
- Consider contextual segmentation: Device type, time of day, and traffic source are signals that can meaningfully differentiate shopper intent. Check whether your current platform incorporates these dimensions.
- Native integrations lower the barrier: The availability of a native Shopify integration means that smaller merchants can evaluate this class of technology without a custom API implementation.
Outlook: Intent-Aware Commerce as a New Category
The $10 million raised by Malachyte and the caliber of its backers signal genuine institutional conviction that real-time intent modeling represents a meaningful evolution beyond legacy personalization approaches. With a Shopify integration already live and an API available for enterprise deployments, the technology is moving from proof-of-concept into mainstream commercial availability.
Looking ahead, Motwani has indicated that the company sees its biggest opportunity in converging merchandising and marketing decisions around the same behavioral intelligence layer. For e-commerce operators who have long managed these functions with separate tools and separate data models, that convergence — if it can be delivered reliably at scale — would represent a significant operational shift in how online stores are run.