We help you define the best architecture in terms of transformation pipeline and storage. For example, we set up ETLs to feed a DataLake and thus centralise the data.
Image recognition
We develop customised image recognition solutions, based on deep learning models and proven open source libraries. Some examples include an OCR solution, a recommendation engine for similar products, anonymisation of sensitive data.
Predictive algorithms
By combining various models such as neural networks, decision trees or SVMs, we implement predictive models. In particular, we have developed models for predicting affluence with applications in revenue management.
Natural language processing
Using LLMs (Large Language Models) such as GPT-4, Llama 2 or Mistral 7B, we implement NLP-based products for feedback analysis, customer service automation and automatic content moderation.
Jean, CEO
Jean has worked at IBM. He is a graduate of the Ecole Polytechnique and EPFL.
Arnaud, CTO
Arnaud has worked at Facebook and Microsoft. He is a graduate of the École polytechnique, from which he graduated at the top of his class, and of the ENS.
We help you define the solution that meets your needs, we implement the algorithms that solve them, we develop the solution, we put it into production and we support you in its evolution.
Some achievements
Projectiv
Application for predicting cinema attendance, developed in-house and resold.
Newsmap
Competitive intelligence tool based on a scoring algorithm for Bandai Namco.
Oulala Deals
Facebook campaign optimization algorithm to maximize revenue.
