Back to Projects
Sustainability, E-commerce

Ango

Built a multimodal recommendation model for a sustainability platform delivered through a Chrome extension, combining visual and textual product signals to generate sub-second recommendations across 10M+ products.

Ango
Global

Key Results

<1s

Recommendation latency

10M+

Products covered

Multimodal

Visual + textual product signals

The Challenge

Helping shoppers make more sustainable choices means surfacing better alternatives instantly, right where they shop — and doing it across a catalogue of more than ten million products. That demands recommendations that are both genuinely relevant and fast enough to feel seamless inside a browser extension.

The Solution

As part of the Halo AI engineering team, we built a multimodal recommendation model delivered through a Chrome extension.

Multimodal Recommendation Model

Combines visual and textual product signals to find the most relevant, sustainable alternatives.

Sub-Second, In-Browser Delivery

Generates recommendations across 10M+ products fast enough to feel instant inside a Chrome extension.

The Impact

Tangible Outcomes

Built a multimodal model combining visual and textual product signals

Generated recommendations in sub-second time

Scaled across more than 10 million products

Delivered seamlessly through a Chrome extension

Key Takeaway

Delivered as part of the Halo team, combining multimodal AI with serious scale to make sustainable shopping recommendations instant and effortless.

Need Recommendations at Scale and Speed?

Let's discuss how multimodal AI can power fast, relevant recommendations across your catalogue.