510 TEAMTERVIEW JORIS WESTERVELD

Joris Westerveld has a MSc in Artificial Intelligence. He first worked as a graduate student and after his graduation he continued to work as a volunteer.

How did you hear about 510?

When I was looking for an internship for my thesis for the master Artificial Intelligence, my mother (who works for the Red Cross since 2014) told me about the 510 and the mission and vision that they have. It instantly peaked my interest, because ever since I finished high school I wanted to do something beneficial for humanity and help where I can, to enhance support systems and agencies. I contacted the 510 through email and made an appointment with Maarten van der Veen (the Strategic Lead of 510). They showed me several potential projects and one of them stood out to me: Predicting Food Security. Being able to create a self-learning model that can be helpful by predicting which areas in a certain country are most likely to transition to food insecurity is highly relevant in this modern age. Moreover, using ‘Big Data’ and Artificial Intelligence to progress the support system in this new age and impact faster and cost effective humanitarian aid is highly relevant and important. Overall the interviews that I had after the email contact with Maarten van der Veen and the other members from the 510 team confirmed my initial thoughts. Namely, it confirmed that I want to help the 510 shape the future of humanitarian aid by converting data into understanding.

What project are you currently working on?

I’m currently creating a model (Xgboost) that can predict whether the food security in a region in Ethiopia improves, doesn’t change or deteriorates. This model only uses variables that are scalable, like satellite imagery, which makes it possible to use this model more easily for other countries. In order to use these scalable variables efficiently, I’ve created a tool that more easily can collect and process satellite imagery.I find it challenging and also highly interesting to use my knowledge about machine learning, statistics (from my bachelor Psychology), optimizing and knowledge of self-learning agents to make this model of food security possible. More over validating this model can also be challenging, since there are important things (like respecting the time line) when validating a time series machine learning model like this.

510’s purpose is ‘Improve speed, quality and cost-effectiveness of humanitarian aid by using data & digital products.’ How are your skills helping 510 reach its purpose?

Using my skills I want to contribute by optimizing humanitarian descision making and aid with regard to food security. In other words my goal is to improve the speed and cost-effectiviness when making these humanitarian decision. In the long run we want to ofcouse also improve the quality of the humanitarian decision making, by for example using this model to get more insight in why and when are these transitions happening.

Comments are closed