In this second Kubeflow article, I build upon what was created in the first one and : Make an advanced pipeline that contains pre-processing, model building, inference and performance evaluation Use GitHub Actions to perform CI and CD Connect GitHub Actions to the Kubeflow endpoint and launch the updated pipeline This article has been peer-reviewed and published in “Towards Data Science”. Kubeflow being new and its documentation being sparse for environments other than GCP made this article a rapid success, with more than 2000 reads in the first 2 days.

In this article, I go through every step that is necessary to have a functioning pipeline : Create a Kuberneter cluster Install Kubeflow Create a container registry Build a container image and push it to your registry Give Kubeflow access to your S3 buckets Create Kubeflow components with input and output artifacts Create a Kubeflow pipeline, upload it and run it This article has been peer-reviewed and published in “Towards Data Science”.

This is an article I posted on Medium that deals with anomaly detection for scoring models’ outputs. Link to the article

This project is my take on the “Real or Not ? NLP with Disaster Tweets” Kaggle competition, in which we are provided a list of tweets that can either be about a disaster or not. All of these tweets have been detected based on keywords such as “ablaze”, and it’s up to the competitor to build a model that determines wheter these words were used in the context of an actual event.

In this project, my goal was to build a predictive model to answer the House Prices Kaggle Competition, in which we are provided with diverse data concerning houses such as quality, number of bathrooms, areas, in order to predict the final sale price. The large majority of the work was spent on exploratory data analysis and feature engineering, with some data cleaning to do in both the training and test set.

A machine learning project in which I used fitbit-like sensor data from gyroscopes and accelerometers to predict the performance of a weightlifting exercise. This project goes through the following key steps : Raw Data gathering and cleaning Model building - RandomForest Cross validation and Testing Link to the GitHub repository