Gerben OostraSemantic Versioning for ML modelsSemantic versioning makes it easy to be explicit about changes in software releases. Could it also benefit machine learning models? Let’s…Mar 22Mar 22
Gerben OostraPolicy Learning without overlap, is it possible?In offline policy learning, one typically requires ‘overlap’ (also known as positivity or common support) for both policy learning and…Feb 1, 2023Feb 1, 2023
Gerben OostraPython Poetry mono repo without limitationsEasily adapting poetry to a mono repo, with released packages and local dev dependenciesJan 9, 20236Jan 9, 20236
Gerben OostraThe future of lead scoring is prescriptiveLead scores aren’t effective in optimizing engagement efforts. A prescriptive approach will improve conversion.Oct 3, 2022Oct 3, 2022
Gerben OostraTo-do list survival guideGetting more and more responsibilities results in different kind of tasks, full agenda’s and less time to focus. As I explained in the…Apr 23, 20211Apr 23, 20211
Gerben OostrainbigdatarepublicTwo steps towards a modern data platformIt’s easy to get lost with many options like Data Lakes, Lakehouses and Data Meshes. How do you go about developing such a data platform?Mar 22, 20215Mar 22, 20215
Gerben OostrainbigdatarepublicUnderstanding inverse propensity weightingAs an ML Engineering craftsman, it’s important to know how models and algorithms work. This allows me to know when approaches work and…Dec 8, 2020Dec 8, 2020
Gerben OostrainbigdatarepublicHow to grow as a data science professional — introducing the Skill StackProfessionals need to grow and develop their skills to advance in one’s career. As a data scientist, it’s no difference. There are a…May 20, 2020May 20, 2020
Gerben OostrainbigdatarepublicPreventing churn like a banditUplift modeling meets causal inference in a bootstrapped reinforcement learning setup.Jan 6, 20202Jan 6, 20202
Gerben OostrainbigdatarepublicFor effective treatment of churn, don’t predict churnSolving the correct problem with machine learning.Sep 13, 20191Sep 13, 20191