Just imagine, it’s a rainy weekend and you feel like watching a film without knowing exactly which one to choose. You open Netflix and start browsing the catalogue. Nothing appeals. There’s so much choice and the catalogue, though well-organised, is so vast that you’re lost. You resort to typing “Best films for a Sunday evening” into the internet and find yourself browsing “must see” selections, “Feel good films” and the top-rated, etc. It’s 11pm and you still haven’t watched anything. Sound familiar? The project by the start-up Sievable may change all that!
Sievable is developing a search engine which enables users to carry out targeted online searches more naturally and effectively. The aim is to no longer drown in a sea of results but instead get access to what you are really looking for.
In practical terms, Sievable is a specializedand collaborative search engine based on machine learning, allowing to very simply carry out specialized research on any domain from a single search engine. The user can filter results according to the user’s own criteria. If we take the above example, the user can access a list that matches their request for“a film”, which opens personalised filters such as“with Will Smith” in the “Science Fiction” category, but “not about robots” and “without aliens”.
This function can be transposed to all fields, such as games or restaurants (“I’m looking for an Italian restaurant within a 10-minute walk from Place de la Bastille, which serves Saltimbocca alla Romana”), but also to the business world and scientific research. For instance, the tool will enable highly specific requests for various research projects from around the world, for example for different teams or for the content of publications, and all via extremely precise and targeted searches. Time-saving is in sight!
Who’s behind Sievable?
Sievable is currently composed of a team of three: Romain Zimmer, Léo Cances and Ragool Krishnan, but they hope to build a future community of contributors to their project!
Romain Zimmer is co-founder of the project. He studied data and theoretical computational sciences at Télécom Paris and applied mathematics at ENS Paris Saclay, with a detour via Deep Learning and computational neurosciences during a research internship at the CNRS in Toulouse. After his internship, Romain joined a company specialised in the production of microprocessors for machine learning, but while he found this experience in the industrial sector interesting, it wasn’t really what he was looking for. He decided to take time out to step back and ask himself some questions about his future and his career, while working independently on the connection between natural language processing and knowledge graphs... which led him to work on search engines!
Léo Cances is the other co-founder of the Sievable project alongside Romain. He holds a Master’s in computer science for aerospace from Toulouse III - Paul Sabatier University and a PhD in artificial intelligence, which he carried out with the Toulouse Computer Science Research Institution (IRIT). He has always been interested in artificial intelligence and has worked on neural network programming since his first year of university. During his studies, he did a work placement in a laboratory on the simulation of the workings of a processor and a placement in a company where he contributed to the design, production and market launch of a product. He enjoyed working in both fields, but in the end the research and innovation aspect was what he found the most appealing.
Ragool Krishnan did a Master’s in software engineering at ISEP, followed by an end-of-study work placement with a start-up in Paris, where he gained his first insight into working in a French tech start-up. He then saw a job offer for a position as web developer at Sievable and was attracted by the interest and originality of the project. He decided to apply for the job and take part in the venture.
A project that has been growing in Paris since May 2021
After doing a benchmark of support schemes for business creation in France, Romain Zimmer opted for the Inria Startup Studio programme. In May 2021, he joined the Startup_Space at the Inria Paris Centre with a project that still needed defining and worked for several months on turning his abstract idea into a concrete project, joined by his business partner Léo Cances.
The support from Inria Startup Studio is going really well.
We always have very interesting discussions, allowing us to obtain different points of view on our ideas and advice on project management and all the different aspects of creating a start-up.
Cofondateur de Sievable
Once the project – the search engine – had been defined, their biggest dilemma was choosing which direction to go in. They decided to create an independent search engine to offer an alternative to those currently available and allow a more natural and efficient way to carry out targeted internet searches regardless of the theme.
At the end of the Inria Startup Studio programme, they hope to find investors and join an incubator.
Today, everyone can be part of the adventure.
Now that the platform has been created, their challenge is to rapidly accumulate enough content and learning data to make the search engine useful. To do so, they are counting on a community of contributors that anyone can be part of. Whether you are interested in machine learning, curious about discovering this solution that could revolutionise internet search codes, or simply want to make your contribution to this Inria entrepreneurship project… anyone can index their content on Sievable!
The three entrepreneurs propose “tokens” that contributors can earn by training the engine and creating content: 1 token per day per item created, with a bonus if the item arrives at the top of the list of search results on the platform.
The tokens will soon be able to be exchanged for other cryptocurrencies.
You can participate in the development of the startup by indexing your content.