We released Recommend UI — the frontend libraries for the new Algolia Recommend product. Recommend is a milestone that made Algolia a multi-product company.
Recommend is a new service that builds machine learning powered recommendations. Two AI models are available: Related Products and Frequently Bought Together.
I led the frontend part of that product; built a team of 4 people, designed the elements, developed the libraries (API client and UI components), and wrote documentation.
I also spent some time researching and developing framework-agnostic UI components while working on it. The idea is to use JSX to describe your UI, and plug your framework's specifics to render this UI description. This final step plugs your framework to our components. They work with any framework that use some kind of Virtual DOM or equivalent: React, Preact, Vue, etc. A bridge to Angular is possible. We implemented this concept in the Recommend UI components, which reduced drastically the code we maintain. This pattern originally emerged when I was designing Autocomplete, where you're able to provide your own renderer. We may write more about it in the future!
Check it out on GitHub and Product Hunt.
Recommend is a new service that builds machine learning powered recommendations. Two AI models are available: Related Products and Frequently Bought Together.
I led the frontend part of that product; built a team of 4 people, designed the elements, developed the libraries (API client and UI components), and wrote documentation.
I also spent some time researching and developing framework-agnostic UI components while working on it. The idea is to use JSX to describe your UI, and plug your framework's specifics to render this UI description. This final step plugs your framework to our components. They work with any framework that use some kind of Virtual DOM or equivalent: React, Preact, Vue, etc. A bridge to Angular is possible. We implemented this concept in the Recommend UI components, which reduced drastically the code we maintain. This pattern originally emerged when I was designing Autocomplete, where you're able to provide your own renderer. We may write more about it in the future!
Check it out on GitHub and Product Hunt.