Datakalab develops Computer Vision algorithms

to better understand people's behaviors

There are more and more people on earth and there is a growing need to understand them

Our DNA : Ethical respect and Protection following GDPR


DATAKALAB analyzes respondents' faces while looking at a content during an online survey.

It delivers an impactful dashboard to better understand attention and emotions, second by second.

Metrics : 

- Attention

- Emotion

- Positive / Negative feelings
- Magic / Pain moments


DATAKALAB improves the understanding and the optimization of customers journey in all places.

We use cameras and mini computers to analyze people's behaviors on device and  we never store any images.

Metrics :
- Attention : stopping power, dwell time

- Analysis of customers' journey

- Emotions : objective Net Promoter Score

- Demographic : Age and gender

- Magic / Pain points


Xavier Fischer

Chief Executive Officer
Co Founder

Graduated from Ecole Centrale Paris.

He was in charge of the European development of Emotient, a company specialized in Artificial Intelligence and Facial Expressions Analysis acquired by Apple in 2016. 

Lucas Fischer

Chief Technology Officer
Co Founder

Graduated from
Berkeley University
and Ecole Centrale Paris 
in Operations and
Data Science.

Lucas is in charge of all Datakalab technologies.

Frank Tapiro

Chief Emotion Officer 
Co Founder

Former advertiser,
creative-director, writer and composer, the less technical but the most emotional.

Founder of "Hémisphère Droit", the Right Brain agency which focused on the brain preferences. He is the inventor of the brands' DNA concept and methodology.

Kevin Bailly

Head of Research

At UPMC, he is a lecturer in Machine Learning since 2011.

Researcher at UMPC, Kevin is an expert in Computer Vision and facial Expressions analysis.

He was in Seoul at ICCV 2019 to showcase his work.

He also won the first prize in 2012 and 2015 of the FERA competition (Facial Emotion Recognition and Analysis) against the M.I.T., Cambridge and UCSD.

Arnaud Dapogny

Ex researcher at Paris Sorbonne University, he works to optimize Datakalab algorithms in order to better understand humans behaviors.

He was in Seoul at ICCV 2019 to showcase his work.

He also won the first prize in 2012 and 2015 of the FERA competition (Facial Emotion Recognition and Analysis) against the M.I.T., Cambridge and UCSD.

Software Engineer

Lucas Lugao

Head of Research

Lucas is Datakalab first employee. He comes from Brazil and graduated from Ecole Polytechnique in software and computer engineering. 

He worked at Google for Verily.

He is the one that thinks and builds each of Datakalab products. 

His goal is to build the most scalable infrastructure for Datakalab.

Alexandre Macedo

Software Engineer

Brazilian by birth, Alexandre graduated from Ecole Polytechnique in Computer Science and Applied Mathematics.

He worked for Thales and is now developing Datakalab infrastructure with passion.

Boris started his career at CEA back in 2012 with the study of the behavior of nuclear fuels under irradiation and plutonium science.

He joined Datakalab to develop light, accurate, and efficient deep learning algorithms to be used in embarked devices.


Boris Dorado

Case studies

OB2A (On Board Advertising Analysis) is the first GDPR compliant product that quantifies crowd audiences. It reveals people attention and emotions in front of outdoor digital billboards to better understand people behaviors when watching a brand content.

It's the first solution that allows the selling DOOH adds by views.

Datakalab analyzes customers reactions during adds on TF1 website.

Knowing  their viewers levels of attention helps the TV channel to optimize its customer experience.

Datakalab analyzes online respondents attention and emotions while looking at a trailer or during a movie in a theater. It allows a better understanding of people's reactions in order to optimize content and to create more impact.

Quantifying the immaterial

Digitalization and robotisation do not always pay attention to human.

Transactional data, digital or relational data are not telling everything  about human behaviors, feelings and reactions.

Datakalab goal is to give a more human face to data.

Our DNA : Privacy by design and ethical respect

As soon as the algorithm detects peoples' faces,

it transforms images into data

and the footage are scratched in less than 100 ms.


Datakalab's goal is to better understand humans without any personal data.

To get more informations, you can consult our Privacy Policy or contact us at

Contact us