![](assets/img/Portrait-CV-transparancy-2.png)
Thomas Bolteau
DevOps Engineer, Data Engineer
GCP Professional Data Engineer
Versatility is one of my main strenghts. Being passionate about engineering, I am always looking for new challenges and technologies.
After my Engineering Degree in Centrale Nantes (France) I decided to study abroad in order to improve my knowledge of electronics and artificial intelligence in Université Catholique de Louvain (Belgium). During that time I also got my GCP Professional Data Engineer certification and worked countless hours on my side projects to improve my skills.
I am currently open to new opportunities in the field of Data Engineering or DevOps all around the world. Feel free to contact me if you have any questions.
![](assets/img/TwitchProfilerV2.png)
Twitch Profiler v2 : Tracking 10M people on Twitch for less than a beer per month
Spring 2023 - Data engineering project
This project is an upgrade of my previous Twitch Profiler v1 project. After getting my GCP Professional Data Engineer certification, I decided to upgrade my project to use the cloud and make a production ready solution.
The last version of the project was definitely not scalable because of the datastructure and the data architecture. Using BigQuery and Cloud storage for the data storage allow scalability. Using Cloud Run to gather the data allow reliability. A major part of the work was to optimize the cost.
A full documentation is available on my Github
In the end I have the daily viewing schedule of more than 10 millions users, with more than 1 millions unique users per day. This architecture would allow me to perform more complex analysis on the data such as NLP analysis on the stream titles to extract highly valuable data.
The system was built using Python, Javascript, BigQuery and many other GCP services.
![](assets/img/techno.png)
Twitch profiler v1 : Tracking 6M user using simple on premise technologies
Summer 2022 - DevOps project
This to create a system to track French users on Twitch streaming website. In the end, I got the daily viewing schedule of more than 6 millions users.
The system was built using Python, MongoDB, Docker, Django and Apache.
A website is available to get a quick look on my data and tools created. More information
Sadly for cost reasons, I had to stop the infrastructure. This project has been updated to a new version with a better architecture and more features.
Machine learning
Knowing how to use TensorFlow in Python and Javascript, I am able to create machine learning models and deploy them on many different platforms.
Here an exemple of digit recognition trained on the MNIST dataset. Feel free to test it.
Draw a number between 0 and 9
Please draw a number