Nathan Nathan van Beelen
Neuroengineer | Interested in ML, Sensors, and Automation
I absolutely love programming—turning ideas into reality with nothing but logic and a keyboard still feels like magic. My journey started with Software Engineering but quickly expanded into Artificial Intelligence, where I explored how code interacts with the brain, processing neural signals and modeling cognition. Neuroengineering in Munich deepened this passion, leading to projects like brain-controlled robots and motion-driven interfaces. After a year of refining my software engineering skills at Sopra Steria, I'm working on a personal project to put my skills further to the test. When I’m not coding, you’ll find me exploring nature, the Alps, or diving into endless curiosities like history, archaeology, languages, and engineering. Curiosity keeps me moving, whether it’s a challenging problem, a new skill, or an adventure outdoors!
Projects
January 2026 - NOW
Building an Autonomous RC Car with an old RC Car and an Arduino
Personal project
In this personal project the goal is to create an RC car that can autonomously navigate and carry out tasks. The brain in this car is the Arduino that replaces the original MCU in the car. It creates environmental awareness with sensors such as ultrasound. The Arduino communicates with the computer over Bluetooth to allow the user to monitor the car and give tasks.
June 2025 - January 2026
Munich, DE
Building a finances overview page for banking consultants
Sopra Steria Custom Software Solutions
While talking to their clients, banking consultants need to have a good overview of their finances. We were responsible tying together different widgets on an overview page to provide this overview to the consultants. My work mainly consisted of writing the back-end to retrieve, transform, and finally provide data to the widgets in the front-end. Later on, I also took over some of the front-end work. Here I focused on the internal logic and writing requests to the back-end.
April 2025 - June 2025
Munich, DE
Building a RAG implementation for QA of company-internal documents
Sopra Steria Custom Software Solutions
The introduction of LLMs brought with it the ability to better search through and digest documents. By building an implementation of a Retrieval-Augmented Generation application, we made it possible for users to upload documents through a web interface for ingestion. The ingested documents could then be used to answer questions asked by the users through the web interface. A major topic I worked on in this project was implementing the batch processing of data in the back-end.
March 2023 - June 2024
Munich, DE
Comparative analysis of zebrafish behaviour with end-to-end learning
Portugus Lab, TUM
Exploratory analysis is the process of analysing a (big) dataset to see if there are salient differences between groups. As this is a virtually inexhaustive, it can make analyses incomplete and biased. By training and analysing Convolutional Neural Networks, I aimed to find a model that can highlight these salient differences between groups. Based on the model, I created a tool that assists in finding quantifiable differences of zebrafish larval behaviour between two conditions, by visualising these differences such that domain experts can interpret them.
October 2023 - May 2024
Munich, DE
Building a microscope for simultaneous imaging and precise optogenetic activation
Portugues Lab, TUM
This project was aimed at building an experimental setup that allows for simultaneous imaging of the zebrafish larvae brain as well as stimulating specific areas using an optogenetic activator. The imaging was done using a light-sheet optical microscope, which allowed for getting a 3D scan of the brain at 2Hz by slicing the brain into numerous slices and taking 2D images. Stimulation of the brain was implemented using a Spatial Light Modulator that is able to reflect light and focus it onto a specific spot in 3D space using digital holography.
September 2022 - March 2023
Munich, DE
Processing of EEG and IMU sensor time data for intention recognition
Institute for Cognitive Systems, TUM
Brain-Computer Interfaces (with EEG) need to be trained individually for every user, and has to be repeated regularly to keep a good performance. This is done by imagining movement, which can often be a tiring and boring task. To ease the training process, I explored the idea of of training with the help of inertial measurement units (IMUs). The main questions to be answered were whether, or rather, how much, this affected the EEG signal quality, and whether the (combination of the EEG and) IMU signals could be used for keeping the user engaged through a game.
March 2022 - June 2022
Munich, DE
Creating an EEG-steered Piano Tiles game
Institute for Cognitive Systems, TUM
Brain-computer interfaces (BCI) translate brain activity to user intent using statistical models that have to be trained on a per-user basis. Not only are our brains plastic, but also the exact conditions of the setup affect the measurements done by BCIs. In practice this means that for a good performance, users have to regularly train the models. This is a rather boring task, even though engagement is important to get good quality data. To counter this, we created a game to keep the user engaged during the training session. Inspired by games like Piano Tiles, the user had to jump between lanes at the correct times to keep playing (and to keep the music going). As the user jumped between lanes by modulating their brain activity, this same brain activity could be used to train the model and maintain a good performance.
October 2021 - March 2023
Munich, DE
Development of a live object-tracking system
Portugues Lab, TUM
The behaviour of larval zebrafish is generally recorded from a whole-arena view. This ensures that the fish is always in view of the camera, but at the same time means that behavioural details are lost, as they cannot be captured by the camera from this distance. In order to get a good view of the fish, but also leave enough space for the fish to behave naturally, we developed an experimental setup where the fish was tracked from a close distance with a camera mounted on linear motors. My focus in this project was to develop and optimise the pipeline from image acquisition to motor correction.
October 2021 - March 2022
Munich, DE
Translating gestures into discrete vibrotactile signals
Msc. Neuroengineering, TUM
This project was aimed at getting hands-on experience with signal processing, effective information transferring, and Arduino. The challenge was to setup the most effective (in terms of the amount of information transfered per second) system of one-way communication between two people. The constraints were that the input had to be real-world gestures for encoding, and the output had to be a number of vibrotactile motors for decoding.
February 2020 - August 2020
Groningen, NL
Evaluating inter-brain synchrony during (dis)agreement in monastic debate
Bsc. Artificial Intelligence, University of Groningen
In my thesis, I quantified the inter-brain synchrony during moments of agreement and disagreement between Tibetan monks during monastic debate. This was done by measuring the EEG signals of both monks simultaneously and annotating camera footage for agreement or disagreement. The moments of agreement and disagreement were quantified in two ways: first a statistical test was done to see if there was a significant difference in synchrony. Then I explored whether it was possible to predict agreement or disagreement on a single-sample basis using machine learning techniques.
January 2020 - June 2020
Groningen, NL
Modelling of fMRI-experiment efficiency
Visual Neuroscience Group, University of Groningen
I developed a simulation program that modelled the efficiency of an fMRI experiment with visual stimuli. By giving the visual stimuli, and the design of the experiment, it gives an estimation of the efficiency of the experiment. This way, the experimenter could already test their experiment design before testing it in an actual fMRI.
Last ~8 years
Setting up and maintaining a Linux server
Personal Project
For the last 8 or so years, I have always had some form of Linux (Debian/Ubuntu) server running. Whether that was a simple Desktop Linux on an old laptop, a manually-built Linux on a single board computer, or a VPS. Over the years I have used it for various purposes, such as email, messaging, web, file synchronisation, backups, and cloud services.
March 2017 - October 2018
Development of a movie/series tracking Android application
Personal Project
I built an Android application to help keep track of movies or series that one watched or is currently watching, including personal details like rating. Rather than storing these details in the cloud, the application stored the user-specific data in a local database, that could be exported in case of a change of device. The application was built on the TMDB API to offer a wide range of shows that the user could keep track of.
June 2016 - September 2016
Assen, NL
Development of the back-end of a quiz web application
IPMarketing
The goal of the project was to develop a prototype for a quiz web application using JavaScript.
September 2015 - June 2016
Assen, NL
Development of the Android and iOS applications for IPMarketing
IPMarketing
The goal was to expand the IPMarketing application to mobile devices by building applications for Android and iOS. I was responsible for writing these applications, including new endpoints in the back-end.
Education
November 2020 - July 2024
Elite Msc. Neuroengineering
Technical University of Munich
GPA: 8.5
Thesis topic: comparative analysis of zebrafish behavior with end-to-end learning.
September 2017 - September 2020
Bsc. Artificial Intelligence
University of Groningen
GPA: 7.6
Thesis topic: Inter-brain synchrony during (dis)agreement in monastic debate.
September 2016 - September 2017
Prop. Software Engineering
Hanze University of Applied Sciences
GPA: 8
Publications and Awards
-
Co-Author in Paper Publication: "A behavioral setup for capturing fine grained coordinated movements of zebrafish larvae"
Preprint Published on Zenodo; March 2025 -
Poster Presentation: "Deep Learning-Guided Analysis of Larval Zebrafish Behavior"
Presented at Bernstein Conference 2023 -
Hackathon: 2nd place with the topic "Flex-Schedule: Dynamic Waiting Lists for Hospital Appointments"
Healthcare Hackathon Bayern 2024
Skills
Programming languages
Tools
Languages
Native
Business fluent
Proficient
Basic