Cube AI brings together an interdisciplinary team with experience across Machine Learning, Physics, and Growth, determined to bring AI to the real world.
Louis de Benoist
Co-founder & CEO
Louis grew up surrounded by railroads and trucks in the logistics hub of Chattanooga TN. Learning to code at the age of 14, he left the US for France to study mathematics and AI.
He majored in mathematics at Ecole Polytechnique, where he worked on gradient temporal difference learning methods for Reinforcement Learning (supervised by Eric Moulines). Then, he studied machine learning as a graduate student at the University of Cambridge. In Cambridge's Computational Morphogenomics group, he adapted a probabilistic framework to develop a generative model which made strides in producing time-efficient tension estimation for epithelial cells (supervised by Bianca Dumitrascu).
Louis worked at OneAngstrom where he developed a new deep learning model for ligand design. He previously founded Hypesetters, an AI platform that optimized online micro-influencer campaigns.
He majored in mathematics at Ecole Polytechnique, where he worked on gradient temporal difference learning methods for Reinforcement Learning (supervised by Eric Moulines). Then, he studied machine learning as a graduate student at the University of Cambridge. In Cambridge's Computational Morphogenomics group, he adapted a probabilistic framework to develop a generative model which made strides in producing time-efficient tension estimation for epithelial cells (supervised by Bianca Dumitrascu).
Louis worked at OneAngstrom where he developed a new deep learning model for ligand design. He previously founded Hypesetters, an AI platform that optimized online micro-influencer campaigns.
Sacha comes from a family of computer scientists, and has been hacking atoms and bits since he was a kid. He studied statistical physics at Ecole Polytechnique.
During his PhD at College de France (supervised by Hervé Turlier), he developed 3D image analysis algorithms for biology based on physical principles. He developed a breakthrough state of the art method to infer forces based on adjoint-state optimisation, published in Nature Methods, highligthed by the CNRS and Nature. With researchers from UCSF, he used Optimal Transport theory to develop a method to measure membrane flow from videos of optogenetically manipulated cells in an article published in the leading biology journal Cell.
He later got interested in GPU-accelerated computing to make his algorithms run faster, and developed his own CUDA kernels to accelerate Fourier Transforms, allowing to generate artificial microscopy data at a 10000X speed. More recently, he adapted the popular LLM inference library Outlines to Apple Silicon chips, quoted by Apple developers in their list of cool projects.
During his PhD at College de France (supervised by Hervé Turlier), he developed 3D image analysis algorithms for biology based on physical principles. He developed a breakthrough state of the art method to infer forces based on adjoint-state optimisation, published in Nature Methods, highligthed by the CNRS and Nature. With researchers from UCSF, he used Optimal Transport theory to develop a method to measure membrane flow from videos of optogenetically manipulated cells in an article published in the leading biology journal Cell.
He later got interested in GPU-accelerated computing to make his algorithms run faster, and developed his own CUDA kernels to accelerate Fourier Transforms, allowing to generate artificial microscopy data at a 10000X speed. More recently, he adapted the popular LLM inference library Outlines to Apple Silicon chips, quoted by Apple developers in their list of cool projects.
Sacha Ichbiah
Co-founder & CTO
Victor Plaisance
Co-founder & CPO
Victor is a hustler with experience in data science and sales. He launched multiple businesses during his master at the first European business school, won regional cycling championships in France, and ranked 17th in the foiling world championships.
He learned to navigate complex enterprise sales cycles while working alongside the CEO of ASO, Yann Le Moenner. Victor was in charge of selling TV and written press media rights for the Tour de France 2021 in over 200 countries, for a total value above 120 million euros. He also led a data science project to optimize broadcaster logistics. Victor pursued data science at Capgemini with Alexandre Teixeira, where he focused on marketing optimization modeling with CAC40 companies, such as L'Oréal.
Later, he founded a social network for high schoolers named Pov, which quickly found early product-market fit in France. The platform reached 50k MAU and quickly expanded into Italy and Spain, driven by a small team of engineers and leveraging a gtm strategy led by high schoolers.
He learned to navigate complex enterprise sales cycles while working alongside the CEO of ASO, Yann Le Moenner. Victor was in charge of selling TV and written press media rights for the Tour de France 2021 in over 200 countries, for a total value above 120 million euros. He also led a data science project to optimize broadcaster logistics. Victor pursued data science at Capgemini with Alexandre Teixeira, where he focused on marketing optimization modeling with CAC40 companies, such as L'Oréal.
Later, he founded a social network for high schoolers named Pov, which quickly found early product-market fit in France. The platform reached 50k MAU and quickly expanded into Italy and Spain, driven by a small team of engineers and leveraging a gtm strategy led by high schoolers.
Cube AI brings together an interdisciplinary team with experience across Machine Learning, Physics, and Growth, determined to bring AI to the real world.
Louis de Benoist
Co-founder & CEO
Louis grew up surrounded by railroads and trucks in the logistics hub of Chattanooga TN. Learning to code at the age of 14, he left the US for France to study mathematics and AI.
He majored in mathematics at Ecole Polytechnique, where he worked on gradient temporal difference learning methods for Reinforcement Learning (supervised by Eric Moulines). Then, he studied machine learning as a graduate student at the University of Cambridge. In Cambridge's Computational Morphogenomics group, he adapted a probabilistic framework to develop a generative model which made strides in producing time-efficient tension estimation for epithelial cells (supervised by Bianca Dumitrascu).
Louis worked at OneAngstrom where he developed a new deep learning model for ligand design. He previously founded Hypesetters, an AI platform that optimized online micro-influencer campaigns.
He majored in mathematics at Ecole Polytechnique, where he worked on gradient temporal difference learning methods for Reinforcement Learning (supervised by Eric Moulines). Then, he studied machine learning as a graduate student at the University of Cambridge. In Cambridge's Computational Morphogenomics group, he adapted a probabilistic framework to develop a generative model which made strides in producing time-efficient tension estimation for epithelial cells (supervised by Bianca Dumitrascu).
Louis worked at OneAngstrom where he developed a new deep learning model for ligand design. He previously founded Hypesetters, an AI platform that optimized online micro-influencer campaigns.
Sacha Ichbiah
Co-founder & CTO
Sacha comes from a family of computer scientists, and has been hacking atoms and bits since he was a kid. He studied statistical physics at Ecole Polytechnique.
During his PhD at College de France (supervised by Hervé Turlier), he developed 3D image analysis algorithms for biology based on physical principles. He developed a breakthrough state of the art method to infer forces based on adjoint-state optimisation, published in Nature Methods, highligthed by the CNRS and Nature. With researchers from UCSF, he used Optimal Transport theory to develop a method to measure membrane flow from videos of optogenetically manipulated cells in an article published in the leading biology journal Cell.
He later got interested in GPU-accelerated computing to make his algorithms run faster, and developed his own CUDA kernels to accelerate Fourier Transforms, allowing to generate artificial microscopy data at a 10000X speed. More recently, he adapted the popular LLM inference library Outlines to Apple Silicon chips, quoted by Apple developers in their list of cool projects.
During his PhD at College de France (supervised by Hervé Turlier), he developed 3D image analysis algorithms for biology based on physical principles. He developed a breakthrough state of the art method to infer forces based on adjoint-state optimisation, published in Nature Methods, highligthed by the CNRS and Nature. With researchers from UCSF, he used Optimal Transport theory to develop a method to measure membrane flow from videos of optogenetically manipulated cells in an article published in the leading biology journal Cell.
He later got interested in GPU-accelerated computing to make his algorithms run faster, and developed his own CUDA kernels to accelerate Fourier Transforms, allowing to generate artificial microscopy data at a 10000X speed. More recently, he adapted the popular LLM inference library Outlines to Apple Silicon chips, quoted by Apple developers in their list of cool projects.
Victor Plaisance
Co-founder & CPO
Victor is a hustler with experience in data science and sales. He launched multiple businesses during his master at the first European business school, won regional cycling championships in France, and ranked 17th in the foiling world championships.
He learned to navigate complex enterprise sales cycles while working alongside the CEO of ASO, Yann Le Moenner. Victor was in charge of selling TV and written press media rights for the Tour de France 2021 in over 200 countries, for a total value above 120 million euros. He also led a data science project to optimize broadcaster logistics. Victor pursued data science at Capgemini with Alexandre Teixeira, where he focused on marketing optimization modeling with CAC40 companies, such as L'Oréal.
Later, he founded a social network for high schoolers named Pov, which quickly found early product-market fit in France. The platform reached 50k MAU and quickly expanded into Italy and Spain, driven by a small team of engineers and leveraging a gtm strategy led by high schoolers.
He learned to navigate complex enterprise sales cycles while working alongside the CEO of ASO, Yann Le Moenner. Victor was in charge of selling TV and written press media rights for the Tour de France 2021 in over 200 countries, for a total value above 120 million euros. He also led a data science project to optimize broadcaster logistics. Victor pursued data science at Capgemini with Alexandre Teixeira, where he focused on marketing optimization modeling with CAC40 companies, such as L'Oréal.
Later, he founded a social network for high schoolers named Pov, which quickly found early product-market fit in France. The platform reached 50k MAU and quickly expanded into Italy and Spain, driven by a small team of engineers and leveraging a gtm strategy led by high schoolers.