Avatar

Dr. Johan von Forstner

Principal ML Engineer

NEC OncoImmunity, Oslo

About me

As a Machine Learning Engineer with a Physics background, I am curious about the intersection of cutting-edge research, advanced algorithms, and real-world data and applications. I am particularly interested in the many ways machine learning can be applied to real-world problems, from perception algorithms processing complex sensor data to solving challenging problems in scientific research. I enjoy tackling the engineering challenges of translating research into robust, practical solutions that operate reliably in complex environments.

This interest in applying ML to challenging real-world problems has led me to my current role at NEC OncoImmunity in Oslo, where I support the development of ML models that are used to design novel vaccines and therapies against cancer and infectious diseases.

My previous experience spans software engineering, machine learning, data science and data engineering across several fields:

As tech lead in the AI team at PARADOX CAT, I worked for 5½ years designing and developing algorithms for automotive in-cabin sensing for high-profile customers such as BMW. Projects combined computer vision and radar technologies and spanned applications such as driver monitoring, occupant detection and classification, or hand gesture recognition, often with real-time or low-power constraints. Some of the algorithms I built are now deployed in production cars.

During my PhD in Space Physics at the University of Kiel, I was part of the scientific teams of the Solar Orbiter Energetic Particle Detector and the Radiation Assessment Detector onboard NASA’s Mars Science Laboratory mission. My research focused on solar eruptions (coronal mass ejections) and cosmic rays. I also contributed to the development of data pipelines and analysis tools for these and other space missions. You can find links to my scientific publications and presentations further down on this website.

Education

  • PhD in Space Physics, 2021

    University of Kiel

  • MSc in Physics, 2018

    University of Kiel

  • BSc in Physics, 2016

    University of Kiel

Blog

Bringing an old rotary phone back to life

How I revived a 1950s rotary phone with a Raspberry Pi Zero.

Experience

 
 
 
 
 

Principal Machine Learning Engineer

NEC OncoImmunity AS

10/2026 – Present Oslo, Norway
NEC OncoImmunity, part of NEC Bio, develops and uses Machine Learning models to design novel vaccines and therapies against cancer and infectious diseases. As a Principal ML Engineer in the Infectious Diseases team, I collaborate closely with bioinformaticians to develop an end-to-end pipeline that models cell and immune system system processes in order to identify and select promising vaccine element candidates.
 
 
 
 
 

AI R&D Architect

Paradox Cat GmbH

07/2024 – 09/2026 Munich, Germany

PARADOX CAT is an established software supplier known for HMI and 3D graphics development services in the automotive industry. As technical lead and and mentor for the new AI R&D team, I specialized in automotive in-cabin sensing with computer vision and radars (60 GHz, UWB).

My team and I supported customers like BMW in both PoC and series development projects across the whole ML engineering workflow from sensor choice, calibration and data collection, through signal processing and training of deep learning models, and all the way to optimization for embedded deployment. Applications of our algorithms include both safety and comfort features based on driver and occupant monitoring in the cabin.

In a series development project for BMW, I developed new signal processing algorithms and edge AI models for Child Presence Detection (detecting children left behind in a vehicle) using UWB radar. My models significantly boosted the performance and efficiency compared to previous approaches and are now released in production vehicles. This success contributed to the 5-star Euro NCAP rating of the BMW iX3 “Neue Klasse”.

 
 
 
 
 

Principal AI R&D Engineer

Paradox Cat GmbH

07/2022 – 06/2024 Munich, Germany
As part of the BETA labs R&D team at our customer BMW, I built prototypes for new in-cabin sensing and natural interaction features in future BMW vehicles, which use computer vision as well as other novel sensors. This involves collecting datasets, implementing data processing algorithms, training Deep Learning models, and deploying these into demo applications running on embedded platforms.
 
 
 
 
 

Senior AI R&D Engineer

Paradox Cat GmbH

03/2021 – 06/2022 Munich, Germany
 
 
 
 
 

Research Associate / PhD student

University of Kiel, Extraterrestrial Physics Department

10/2018 – 02/2021 Kiel, Germany

The Space Physics group at Kiel University designs and operates their own instruments to measure space radiation aboard various ESA, NASA and CNSA missions and analyzes their data to advance solar and planetary sciences as well as astronaut safety.

I served as lead instrument scientist for the High Energy Telescope onboard ESA’s Solar Orbiter and as a member of the research teams for the Mars Science Laboratory (“Curiosity” Mars rover) mission as well as Chang’E 4, China’s mission to the far side of the Moon. My research on solar eruptions and their interaction with cosmic rays resulted in 25 peer-reviewed papers, including four as first author.

In addition to my research, I developed and maintained Python-based data pipelines and analysis tools for multiple instrument teams, contributed to several open-source libraries (including matplotlib, pandas, sunpy and SpiceyPy), and taught lab courses in experimental physics.

 
 
 
 
 

Student Research Assistant

University of Kiel, Extraterrestrial Physics Department

08/2016 – 10/2018 Kiel, Germany
 
 
 
 
 

Software Developer

rami.io Software Development

08/2014 – 02/2021 Heidelberg, Germany (remote)

rami.io (now pretix GmbH) is a young, bootstrapped company behind open-source projects such as the pretix event ticketing system that grew from 1 to 10 employees during my time there.

I started as an Android developer on the Web Opac App project, developing new integrations to support more than 1,000 public libraries around the world. Later, I also served as lead developer for the Vertretungsplan.app project, building the backend as well as Android and iOS apps to notify high school students and teachers of schedule changes in real time. Besides these two main projects, I occasionally supported development and customer support for pretix and its Android companion apps for ticket sales and scanning.

Open Source Projects

that I started or contributed to

car-templates-host

Android Automotive OS template host with map implementation.

EVMap

Android app to find electric vehicle charging stations.

Web Opac App

Android App to access library catalogs

Orbit Plots

Artistic plots of Solar Orbiter data

GCS Python

Python 3 implementation of the Graduated Cylindrical Shell model for coronal mass ejections

substitution-schedule-parser

Java library for parsing schools’ substitution schedules

Presentations

UWB to the rescue – cost-effective in-cabin sensing with AI and sensor fusion
First CME and Forbush decrease observed at Solar Orbiter using EPD

Publications

Multipoint Interplanetary Coronal Mass Ejections Observed with Solar Orbiter, BepiColombo, Parker Solar Probe, Wind, and STEREO-A (2022)

Evidence for local particle acceleration in the first recurrent Galactic Cosmic Ray depression observed by Solar Orbiter. The ion event on 19 June 2020 (2021)