At Joko, we believe that today’s online shopping experience is fundamentally flawed. How many times have you found yourself juggling between dozens of tabs to find THE right product? How many times have you felt completely lost in a jungle of information? Let's face it, the online shopping experience can be extremely frustrating: comparing products, having to navigate through confusing menus and filters, keeping track of prices etc. quickly becomes a nightmare. And so-called intelligent search services like Google Shopping are often useless.
At Joko, we are convinced that this cannot be the future of online shopping, and we are putting a lot of effort into revolutionizing it. We are crafting a product that enables users to find their desired products in the smoothest way possible, to effortlessly compare all their characteristics, and to obtain transparent and clear information on both their price and environmental cost. Today, Joko is behind an app in the top 10 most downloaded shopping apps in France, with 3 million users, and it's rapidly expanding in the United States.
We are a tight-knit team made up of profiles from the best universities (Berkeley, Polytechnique, Cambridge, ENS, Mines, MIT, etc.), and doctors from the best research labs, as AI/ML algorithms play a central role in what we are building. We value talents from diverse backgrounds and experiences to create great things together.
First of all, to achieve our goal, we are building the world's largest product catalog, a universal catalog composed of all the products sold by all e-commerce sites in the world. For this, we need to understand any web page and extract clean and structured information from it. We have developed LLM-based approaches to address this problem. One of the major challenges is to scale these approaches on colossal volumes of data (we have to process hundreds of millions of pages several times a day). We have developed state-of-the-art approaches, but there is still a lot of research needed to optimize their performance and resource efficiency.
⇒ More information in 📰 Web Pages and LLMs, a Match Made in Heaven
Another significant challenge is enhancing and structuring our product catalog to maximize its potential. For example, we need to identify identical or similar products, categorize them as accurately as possible, detect and discard fake information, to automatically link product models (e.g., iPhone 15) to all their variants (e.g., blue, pink, 128GB, 256GB etc.), etc. Here again, one of the major challenges is dealing with gigantic volumes of data with language models and vision models.
Finally, we want to understand the desires and intentions of users to develop the next generation of shopping assistants. Again, recent advances in the field of GenAI allow for the development of very powerful approaches, but a big challenge is to develop always reliable approaches, with low latencies and controlled costs. This is another area in which we invest a lot of resources at Joko today.
⇒ More information in 📰 The Future of E-Commerce: Shopping Online with your AI Assistant
Joko has been offering annual research internships in Machine Learning for several years. All our internships are closely tied to our engineering teams to maximize their tangible impact.
In 2024, we wish to explore the following research topics:
We generally offer internships adapted to both the will of the candidates and the potential for Joko. Contact us for more details.
You can contact the following people who did a ML research internship at Joko and joined us full-time after their internship.