Leader Data Tribe

• Feb 2023 - Present

TFP consists of numerous tribes and as leader of the Data Tribe I coordinate our initiatives and meetings. Also, in preparation of TFPs quarterly meetings, I promote our interests towards management. As Data Tribe we currently are 7 data specialists and are always looking for passionate data fanatics.

Data Scientist

• Mar 2022 - Present

The Future Partners (TFP) is a collective of ±50 IT freelancers. The purpose of TFP is to share our experiences and knowledge, to grow and to have FUN.

Senior Data Scientist - Computer Vision

• Jan 2023 - Present

Senior Data Scientist - Computer Vision

• Apr 2022 - Jan 2023

PAVEMENT SEGMENTATION OF AERIAL IMAGES 𝐏𝐫𝐨𝐛𝐥𝐞𝐦: Krinkels B.V. is a contractor focused on green maintenance. To estimate the workload and cost of maintaining a particular area, it is vital to know the types and amounts of pavement that resides therein. 𝐑𝐞𝐬𝐩𝐨𝐧𝐬𝐢𝐛𝐢𝐥𝐢𝐭𝐢𝐞𝐬: Jorrit was solely responsible for the delivery of an image segmentation application: from data gathering and pipelining to model development, UI definition and deployment. Jorrit also coached more junior colleages in using Git, performing code-reviews and working with Convolutional Neural Networks. 𝐎𝐮𝐭𝐜𝐨𝐦𝐞: Jorrit delivered Streamlit apps that Krinkels employees can use to (1) train and evaluate new models and (2) make pavement predictions over custom areas. 𝐓𝐞𝐜𝐡-𝐬𝐭𝐚𝐜𝐤: Python, Git, Tensorflow/Keras, U-Net, Streamlit, GeoPandas, RasterIO Show less

Senior Data Scientist @ Rabobank Model Monitoring

• Jul 2021 - Feb 2022

The Rabobank uses Credit Risk models to predict the riskiness of loans and to determine provisions. The adequacy of these models need to be closely monitored and backtested. This is time-consuming work that can be repetitive. I was one of three core developers of a Python software product that automates quantitative parts of the monitoring process. Depending on configuration, it runs specific statistical tests and stores the results in a SQL database living in Dataiku, in a Word document and in HTML format. Furthermore, Jorrit was solely responsible for executing two monitoring projects of valuation models. The software product is delivered as a Python package that is made available via the Rabobank package manager. Fully documented using documentation generator `pdoc`. Furthermore, I enhanced the data quality of monitoring data by organising three-way collaboration between the Business, Data Delivery and Monitoring.𝐓𝐞𝐜𝐡-𝐬𝐭𝐚𝐜𝐤: Python, SQL, Git, Azure DevOps, Azure Pipelines, Dataiku, package-development, pdoc documentation generator Show less

Medior Data Scientist - Computer Vision

• Feb 2021 - Jul 2021

With a small team (1.8 - 3 FTE) we developed an image classification tool specific for diatom images. Diatoms are unicellular algae that are indicative for their environment and therefore function as proxies for water quality and environmental change. Besides classification of images, the tool allows users to browse through our database of images and meta-data regarding the different species. To help the user understand how our model comes to a prediction we have implemented an Explainable AI layer called LIME. The Python Flask backend serves the React frontend via APIs and a MongoDB database. The solution is deployed in a virtual machine and available on the platform 890 by Capgemini.I am the driving force behind the development of all of these aspects since I am the only full time resource. I have presented our results to the CEO of Capgemini I&D Global (counts 18.000 employees). Show less

Medior Data Scientist - Team lead - Computer Vision

• Dec 2020 - Feb 2021

I lead a team of three with whom I developed software that summarizes soccer matches in an automated way. To determine the highlights of the game we make use both the audio and video data.

Junior Data Scientist / Data Engineer @ ING Nederland

• Aug 2020 - Dec 2020

In a difficult stakeholder landscape, we developed a knowledge graph product to connect multiple data sources in the Wholesale Banking domain. I created a synthetic data set that follows the same distributions as the actual production data. This allowed us to develop a prototype without having access to production data. Furthermore, I was responsible for assessing data quality of different sources. For this I used Python (PySpark, Pandas, etc.) and SQL. I addressed the issues to the different data owners while maintaining their faith in our product. Finally, I assisted with the ontology development and the data pipeline (mapping data from a relational database to RDF). Show less

Junior Data Scientist / Software engineer @ Governmental institution

• Sep 2019 - Aug 2020

With a team varying between 4 and 10 people we build an product that verifies the authenticity of identity documents. We build this product from scratch, in Python. The product consists of 12 microservices, each performing a step in the pipeline. Examples of steps are: classification, image preprocessing, text extraction using Tesseract (OCR), verification of textual data and verification of graphical data. The microservices are exposed using Flask. The product is build in an object-oriented way. I was responsible for the microservices that performed face matching, (fuzzy) string matching and multiple graphical verifications. Furthermore, I implemented error handling and he contributed to the design of the software architecture and the database model.This product won the Most Innovative Solution Award in the yearly, worldwide award-session of Insights & Data Capgemini (~15.000 employees). Show less

Onderzoeker / Entrepreneur

• Jan 2017 - Jun 2019

Onderzoek doen naar een nieuwe methode om het droge ogen syndroom te diagnosticeren. Naast het coderen van een data-analyse script heb ik de verantwoordelijkheid op mij genomen om het project te overzien en de grote lijnen uit te zetten. Verder ben ik contactpersoon van verschillende externe partijen en leid ik de bijeenkomsten die met hen gehouden worden. Op https://ibb.co/album/iDkOGF vindt u een selectie foto's die de werkzaamheden van ons team illustreren. Onderzoek doen naar een nieuwe methode om het droge ogen syndroom te diagnosticeren. Naast het coderen van een data-analyse script heb ik de verantwoordelijkheid op mij genomen om het project te overzien en de grote lijnen uit te zetten. Verder ben ik contactpersoon van verschillende externe partijen en leid ik de bijeenkomsten die met hen gehouden worden. Op https://ibb.co/album/iDkOGF vindt u een selectie foto's die de werkzaamheden van ons team illustreren.

Bijlesdocent

• Aug 2011 - Jun 2019

Student-assistent

• Feb 2014 - Jan 2018

Geven van/assisteren bij cursussen Quantumfysica en Elektriciteit & Magnetisme voor middelbare school docenten (https://youtu.be/pHbejALvFwY); Geven van werkcolleges aan eerstejaarsstudenten. Geven van/assisteren bij cursussen Quantumfysica en Elektriciteit & Magnetisme voor middelbare school docenten (https://youtu.be/pHbejALvFwY); Geven van werkcolleges aan eerstejaarsstudenten.

Vicevoorzitter

• Mar 2015 - Mar 2016

Portefeuille: Acquisitie Resultaat vergeleken met voorgaande jaren: €17.000 aan sponsorinkomsten waar voorheen €5.000-€10.000 regulier was: https://ibb.co/album/miDAwF. Portefeuille: Acquisitie Resultaat vergeleken met voorgaande jaren: €17.000 aan sponsorinkomsten waar voorheen €5.000-€10.000 regulier was: https://ibb.co/album/miDAwF.