How Malachyte is Building Behavior Intelligence for "Cold Start" Retail

Malachyte, the first e-commerce AI Platform that adapts in real time, responding to consumer behavior instantly, led by Ian Anderson, Siddharth Motwani, and Shivaditya Sinha, has raised $10M in seed funding, co-led by Bessemer Venture Partners and Gradient Ventures, with participation from Harpoon Ventures, to bring this vision to life.

Before founding Malachyte, co-founders Ian Anderson and his co-founder spent a decade at Spotify working on personalization. Their work included building recommendation systems behind Discover Weekly, improving search and discovery across the platform, and developing dynamic merchandising at scale.

The experience exposed limitations in traditional personalization methods, which were largely manual, unable to adapt in real time, and difficult to scale across Spotify’s 800 million users and more than 1 billion products. The team identified a need for new infrastructure capable of tracking user preferences and predicting intent in real time based on behavior. The approach, described as “behavior intelligence,” was designed not only to address personalization at Spotify but also to enable new applications across a broader range of digital experiences.

Learning From History: From Steel to Intelligence.

The company’s thesis draws on the transformation of steel during the Industrial Revolution. In the early 1800s, steel was expensive, brittle and difficult to produce, limiting it largely to small-batch, niche applications. The Bessemer process changed that by making high-quality steel affordable at scale, unlocking infrastructure such as railways, skyscrapers and modern cities.

Malachyte sees a similar infrastructure shift emerging around real-time intelligence. Its approach focuses on AI infrastructure capable of processing large-scale data and understanding live behaviors and signals to power discovery, search and personalization. While much of the AI industry’s attention is focused on LLMs and agents, the underlying infrastructure that determines what systems understand about user intent, and how quickly they can act on it, could represent a broader technological shift.

The Attentionless Economy.

Modern brands face growing challenges in connecting with customers as traditional digital tools such as search, recommendations, and advertising become increasingly repetitive and rigid. For e-commerce, the result is often a standardized shopping experience built on assumptions that fail to reflect how consumers actually behave.

A more adaptive model would allow brands to understand a shopper’s current intent, interests and needs, and adjust the experience in real time as behavior changes. Most existing personalization technologies, however, were designed for the “logged-in web” and rely heavily on past purchases and predefined rules, despite much of today’s traffic being anonymous.

This can lead to inaccurate assumptions. A shopper who previously purchased an infant onesie, for example, may be classified as a new parent even if other behavior provides no supporting signal, the purchase could simply have been a gift for a colleague.

From Recommendations to Behavior Intelligence.

In the attention economy, relevance depends on live intent, what shoppers signal in the moment, and how those signals evolve over time. Like the Bessemer process transformed steel production, modern personalization requires new infrastructure built around real-time behavioral intelligence.

Malachyte’s approach is designed to rank current intent rather than rely on stale purchase history, provide a transparent merchandising layer instead of an opaque ad auction, and remain modular enough to be continuously tested and adjusted.

The company calls this approach “behavior intelligence,” using continuously updated models to understand user intent across search, recommendations and product pages. Unlike legacy systems that rely on batch retraining, rules and surface-specific models, Malachyte’s infrastructure is designed for continuous learning, rapid adaptation and real-time performance at scale.

By combining behavioral data with continuous learning infrastructure, Malachyte aims to enable digital experiences that adapt proactively to changing shopper needs while helping brands and retailers maintain performance as consumer intent evolves.

A Vision for Modern Commerce.

Modern personalization is moving beyond static purchase histories toward a continuously updated understanding of consumer intent. Malachyte’s behavior intelligence infrastructure learns from real-time signals such as clicks, hovers, comparisons, and abandoned purchases, allowing digital experiences to adapt as shopper behavior changes.

The approach is designed to reduce reliance on historical data, predefined rules, third-party cookies and frequent model retraining. By learning continuously from current behavior, the system aims to deliver more relevant predictions without requiring extensive user histories.

For consumers, this translates into product discovery that evolves with curiosity, search that better understands intent rather than just keywords, and recommendations that respond to context rather than relying on generic similarity models.

For brands, the infrastructure enables digital experiences to adapt to each visitor without requiring extensive prior data or manual rules. Rather than functioning as a chatbot, Malachyte positions behavior intelligence as an underlying real-time layer for discovery, search, and personalization, using vector-based AI techniques to interpret user behavior and intent.

Built by a Team with Frontier Experience.

Malachyte was founded to move beyond AI systems that reduce consumers to clicks and broad behavioral cohorts. The company’s approach focuses on understanding momentary intent without relying on extensive identity data, while giving brands greater transparency and control over how personalization decisions are made.

Co-founders Ian Anderson and Shivaditya Sinha, alongside a team of AI specialists, bring experience spanning large-scale systems, consumer behavior, and major commerce operations. The company is also expanding its customer base among brands and retailers seeking to rethink personalization and product discovery and create more adaptive shopping experiences.

𝐑𝐞𝐦𝐨𝐭𝐞 𝐇𝐢𝐫𝐞 𝐖𝐢𝐭𝐡 𝐔𝐬: 𝐌𝐞𝐧𝐥𝐨 𝐓𝐚𝐥𝐞𝐧𝐭 helps technology companies hire exceptional remote talent from India. Build your team with carefully curated engineers, product leaders, designers, GTM professionals, and more. Learn More At: https://www.menlotimes.com/menlo-talent