Testimonial Volunteer Research Assistant / Data Scientist for the Dessine Moi La High Tech Association

 

 

Presentation of Dessine Moi La High Tech Association

The association is divided into 2 components: patient support and research aid.

  • Patient support : through its actions, Dessine-moi la High-Tech wishes to improve the comfort and mental health of hospitalized children and their families. In fact, the hospitalization conditions of children with cancer, especially those hospitalized in a protective isolation room, impose certain constraints that considerably limit their possibilities for distractions and escape. The objective of the activities offered by the association is therefore to allow children and their families to enjoy new, fun and entertaining experiences, allowing them to forget, for a few moments, the hospital context.


Offer moments of escape to sick children

Dessine-moi la High-Tech - Nos ateliers à Nantes

 

  • Research Aid : Along with the high-tech workshops and in line with its # TechForGood * strategy, the association's ambition is to support research into pediatric cancers. How? 'Or' What ? By offering human assistance to researchers working on pediatric cancers using artificial intelligence (AI) tools with the provision of volunteer Data Science Experts from the world of industry. In recent years, in medicine and more particularly in cancerology, genomic and molecular data (DNA, RNA, proteins, etc.) have been available in large quantities which must be analyzed by sophisticated computer tools. The contribution of Artificial Intelligence and Machine Learning is essential in the analysis of this large-scale data: these tools make it possible in particular to highlight signatures and potential biological mechanisms within this mass of data that cannot be understood by the only human brain. In oncology in particular, it is thus possible to individualize molecular diagnostic subtypes, and to identify factors of prognosis and prediction of response to treatments using these tools, allowing better, personalized care of the patient with cancer.

 

If you want to know more about the association, please visit this website : https://www.dessine-moi-la-high-tech.org/ 
Make sure to discover the whole podcast!


 

Who are we ?

Zeineb Ghrib: I have been a data scientist (junior) at Prevision.io for 2 years, basically I have an engineering degree in telecommunications but I have changed my vocation.

I love philosophy, history, manga and machine learning, but above all I really like to feel useful ..
 

Mathurin Aché: I am 40 years old, father of 3, soon to be 4 children. I have been a data scientist for 17 years, I work for the start-up Prevision.io. I like numbers, new technologies, sports and strategy games.

 

How it all began ?

Mathurin : It all started after the summer of 2020 through a telephone exchange with Jérôme Roul who introduced me to the Dessine Moi La High Tech association and who thought that I could be a good contributor. I had known Jérôme from my previous professional experience at Orange. He presented me with the missions of the association's Research Pole as well as the content of his initial discussions with doctors, and told me how we, data scientists, could help them.

Very quickly, I liked the subject and I got involved in a first project with the Institut Curie in Paris: on the one hand, to prove myself and on the other hand, to see if we could really bring them a useful contribution.

The interactions with our interlocutor at the Institut Curie were very constructive and very motivating.

And it was quite natural that I started talking about the association around me, to my knowledge of my Data Science network, and I convinced my colleague Zeineb to join us on this project.

 

Zeineb : It was Mathurin who encouraged me to participate in activities outside of work. I always thought that with my level experience I could not contribute much, but Mathurin gave me confidence, and above all I wanted to participate in non-profit purpose projects even as a junior.

In addition, I noticed that there are a lot of people invested from different levels and fields which encouraged me a lot. I am very happy to have brought or tried to bring something, even a little, it is always encouraging!

 

In a nutshell, our experience as a volunteer consists of

Mathurin

  • Become aware of the difficulties encountered by doctors, researchers, and other specialists in medical data, concerning pediatric cancers,

  • Translate these issues into data science approaches: What data can I have? Which metric should I optimize? What algorithm can I use? Is there a reference solution? What validation strategy should I suggest? Is understanding and interpreting the results important? How to share the predictions? How often should I plan for retraining?

  • The questions we are trying to answer relate to better identification of the type of pediatric cancer, and we are using data from RNA sequencing. 

Zeineb

  • Discuss with people in the business, experts in the field of cancer research to learn about the problem and translate it into data science terms.

  • For me, the work of Mathurin was already a high level baseline, at the beginning I tried to improve the existing performances with a new type of model or another evaluation method (CV) .. but in vain, which led me to tell myself that instead of improving Mathurin's work I could supplement it: by giving an explicability to the existing model and thus avoiding the “Black Box” mode of the AI: Indeed I wanted to give a local interpretation (for each prediction made) of the model which can be very useful for a doctor. By breaking down the predictions, it can be shown that this type of method can provide very effective and relevant alternatives to help specialists make more detailed diagnoses and both classify and interpret patient records.

  • Exposing your work to expert people in oncology and finding that the analysis of the modeling results is consistent and very speaking to them is very encouraging. In another words you have provided an added value, without having any medical knowledge

 

Any fears upstream ....

Zeineb: I was afraid that my contribution was insignificant or that I didn't have the right skills to participate in the project.

Moreover, I have recently been given new responsibilities in my personal life, so I did not want to get involved without ensuring afterwards, but Mathurin confirmed to me that our company agreed that we participate during our working hours.

In addition, I thought that we would have medical data accessing issues which would be constraining for the project, however, using public data was a good alternative.

In terms of resources, and due to the possible risks of using personal data, we were unable to use more powerful cloud servers. But this constraint does not concern public data that we used for this project
 

Mathurin: When I joined the association, I had several fears:

  • That my company does not accept that I contribute to this association during my working hours. But on the contrary, Prevision.io was very pleased with our initiative to participate in data science missions with an associative objective, and supported us in this project which is particularly close to our heart and which, indirectly, contributes to the influence of society, beyond the companies that have placed their trust in us.

  • That the association thinks that this is a commercial move of Prevision.io and that I cannot, therefore, participate.

  • Not having the necessary skills to help researchers develop their knowledge of pediatric cancers.

 

What does it bring us (humanly, intellectually, ...) ?

Mathurin: By participating in these missions, I have the feeling:

  • to be useful by contributing at my level to a better knowledge of rare diseases,

  • to participate in a human, social project where the end is not financial gain, but in which the main objective is the children sake,

  • to work with other committed and invested data scientists,

  • learn with experts in the medical field and acquire new skills,

  • From a slightly more individual point of view, by taking an active part in this association, I feed my CV with a great project and expand my network.

 

Zeineb: As a data scientist, contributing to a project for a non-profit purpose, different from the business projects that we used to work on such as: fraud detection, retail, targeting, bank's grant score ..., is very rewarding from a moral point of view.

In addition, I was able to meet very competent people who are very invested in the field of medical research with whom we shared very fruitful experiences.

But also, this teamwork allowed me to expand my knowledge and acquire new skills while having the feeling of being useful and bringing an added value to the project.

 

And an indirect consequence is that this contribution will remain a valuable asset in our CV

 

Conclusion

Zeineb:  Thank you for having accepted me as a member of the association, and I would like to express my respects to Jérome Roul, Gregory, Camille and all the members of the association who work for free to advance research in the field of cancer in children.

But I also want to say that I am pleasantly surprised that France has such resourceful and committed people in this area.

I remain convinced that the involvement of the AI aspect in this field of research constitutes a great advance and will accelerate its progress.

In addition to what has already been done, it would be a good idea to work even more on communication, so that the work of the association receives a wider echo: rebroadcast on this type of podcast on several channels, publications, presentations of work on online forums / meetups.

 

Mathurin: 

  • I would like to thank Grégory and all the members of the association for their struggles to improve the lives of children with cancer, it is a chance to be able to contribute with you.

  • I also wanted to thank Jérôme Roul who from day one has been involved in growing the research component of the association.

  • To all the people who ask me how they can contribute to the association:

  • I invite you to relay the association's messages, especially on social networks,

  • If you are a data scientist, and the idea of ​​contributing to research in order to improve the diagnosis / care of children suffering from cancer, come join us, the team is welcoming, and the cause is as stimulating as just.


     

 

ABOUT PREVISION.IO

Prevision.io brings powerful AI management capabilities to data science users so more AI projects make it into production and stay in production. Our purpose-built AI Management platform was designed by data scientists for data scientists and citizen data scientists to scale their value, domain expertise, and impact. The platform manages the hidden complexities and burdensome tasks that get in the way of realizing the tremendous productivity and performance gains AI can deliver across your business.

 


Suivez-nous sur Twitter ! Suivez-nous sur Linkedin ! Suivez-nous sur Instagram !