I'm finishing my PhD at the intersection of NeuroAI and Computational Cognitive Neuroscience at the Max Planck Institute in Leipzig, where my research focuses on datasets used for comparing neural network models with the human brain.
Before academia, I spent several years working on applied data and ML problems - building recommendation systems, image processing pipelines, and ML infrastructure in production. What drives me is solving hard technical problems that actually matter: whether that's making neuroimaging experiments more efficient or providing a model that changes how a product works.
Hello from the other side! If you're here, it means you want to learn more about me personally - so I'll keep things more casual!
Here you'll find some info about me and my hobbies, my taste in music, movies and books, and some notes on things I'm thinking about.
Hebart Lab · Max Planck Institute for Human Cognition and Brain Sciences & University of Gießen
Researching NeuroAI approaches for efficient experimental design in visual neuroscience: using modern computer vision models and large-scale naturalistic image spaces to choose better stimuli, compare model and brain representations, and improve fMRI dataset coverage. Built the ML/data infrastructure behind these workflows, including public datasets and re:vision challenge infrastructure.
Python · PyTorch · CLIP · scikit-learn · SLURM · Docker
ScaDS.AI Dresden/Leipzig
Engineered a multi-plane UNet++ ensemble for Glioblastoma segmentation (BraTS 2021), reaching 0.90 Dice for whole tumor. Developed attention-based mortality prediction models with epistemic uncertainty estimation.
PyTorch · FT-Transformer · SAINT · UNet++
CHECK24 (Travel Vertical)
Designed and deployed a Flask + Redis image-processing microservice for fast ML inference over 20M+ hotel images. Built deduplication, retrieval, classification, quality scoring, recommender tuning, and Grafana monitoring workflows.
Python · PyTorch · Flask · Redis · Hyperopt · BigQuery · Grafana
Kimetric UG
Implemented two academic websites using Django, Nginx, and Gunicorn. Configured Linux hosting environments and automated deployment scripts.
Django · Nginx · Gunicorn · Linux · CI/CD
Webdata Solutions GmbH (now Vistex)
Revamped a product-matching pipeline with a neural-network approach trained on self-collected web-scraped datasets, increasing matching accuracy from <50% to 92%. Built data collection, training, evaluation, and interpretability workflows.
Python · TensorFlow · PostgreSQL · AWS
Leipzig University
Graduated with grade 1.2 (Distinction). Focused on ML, data analysis, and medical image processing. Thesis on using GANs to synthesize images that maximally activate specific brain regions.
Machine learning · Data analysis · Medical image processing
Leipzig University
Built a foundation across computer science, data systems, software engineering, and business-facing analytics.
Software engineering · Databases · Analytics
Vision neuroscience runs on large fMRI datasets, but nobody had checked whether the stimulus images in these datasets actually cover what humans see in the real world. We built LAION-natural -a reference distribution of ~120M naturalistic photographs filtered from 2 billion LAION images using a CLIP-based classifier trained on 25k actively sampled labels. Then we measured coverage: ~50% of the visual-semantic space is missing from the two most widely used datasets (NSD and THINGS).
The good news: you don't need millions of images to fix this. In both simulations and real fMRI data, out-of-distribution generalization saturates at 5-10k samples - as long as you draw them from a diverse enough pool. We compared seven sampling strategies (random, stratified, k-Means, Core-Set, effective dimensionality optimization, active learning) and found that pool diversity matters far more than which algorithm you use to sample from it.
The pipeline processes billions of images using CLIP embeddings, Annoy indices for nearest-neighbor search, mini-batch k-Means clustering, and Ridge regression encoding models - all at a scale that runs on a university HPC cluster, not a cloud budget.
Most scientists learn to code informally - picking things up as they go, optimizing for "does it run?" over "will anyone else understand this?" This paper introduces a structured framework for writing better research code, built around the idea that researchers naturally switch between quick prototyping and careful development - and that being deliberate about which mode you're in makes all the difference.
The ten principles span three tiers: organizing code (standardized project structures, version control, automation), writing reusable code (testing, documentation, clean interfaces), and collaborating (code review systems, shared knowledge bases, lab-wide standards). Already at 22k+ accesses, it clearly hit a nerve - these are problems every computational lab deals with but rarely talks about explicitly.
You can't show the brain every possible image, so how do you figure out what a specific patch of visual cortex actually responds to? We trained a convolutional neural network end-to-end on fMRI data from a subject watching naturalistic movies - no ImageNet pretraining, just raw stimulus-response pairs. Then we used BigGAN to synthesize images that maximally activate individual voxels via gradient ascent through the model.
Early visual areas (V1-V3) preferred gratings in small receptive fields, as expected. More interesting: FFA showed preference for faces but also oval shapes and vertical symmetry, while PPA preferred places plus horizontal lines and high spatial frequencies. An SVM classifier could distinguish FFA vs. PPA preferred stimuli from their GAN latent vectors at 87% accuracy, confirming the approach produces meaningfully different outputs per region. This was one of the first demonstrations of GAN-based preferred stimulus synthesis for the human visual system.
Dataset notes
Vision research needs naturalistic photographs, but web-scraped datasets like LAION are full of screenshots, memes, ads, and generated images. We scored all 2.1 billion images in ReLAION-2B for "naturalness" using a CLIP-based classifier, then extracted and published ViT-H/14 embeddings for the ~500M most photographic ones. The result is a 167GB dataset on Hugging Face that lets researchers query half a billion images by visual similarity without downloading a single pixel.
Library notes
An open-source Python toolbox for extracting and comparing image representations from deep neural networks. Supports 100+ models across torchvision, timm, CLIP, self-supervised models (DINO, MAE, SimCLR), and more. Also includes tools for aligning DNN representations with human similarity judgments via RSA and CKA. I'm the third-largest contributor to the project, which has 460k+ PyPI downloads and is used across vision and cognitive neuroscience labs.
Dataset notes
LAION-fMRI is a deeply sampled 7T fMRI dataset for studying human visual representations with naturalistic images. It contains brain responses to 25,052 LAION-derived images across 5 subjects and 165 acquired scanning sessions, including single-trial GLMsingle betas, retinotopy, localizers, and diffusion data. The dataset supports work on model-brain comparison, stimulus selection, and replication through the re:vision initiative.
Recognition / service
My love for moving my body started with joining a breakdance crew at the age of 17. Since then I've dabbled in all kinds of movement disciplines. Today, I really enjoy bouldering, but I also do resistance training, yoga and more recently Qi Gong (good way to wake up the body in the morning!). I also love hiking in nature (and combining it with climbing is the best way to spend a day, see above).
A couple of years ago, I asked a bunch of friends, "Hey, want to start a band?" To my surprise, everyone said yes! We had to drag our front singer into it, but now he regularly screams his heart out. We mostly play pop punk covers for friends' birthday parties, but the dream of writing our own songs is still alive and kicking.
As a kid, one of my teachers would say "Johannes lives off plain air", because I refused to eat food from the school kitchen. Either I was very picky, or the food was just plain bad. Today I eat almost anything - but I love throwing together a plate of whole food ingredients into something delicious. Lately, I enjoyed hosting friends for dinner nights, which have been a blast (but surprisingly stressful!).
I've been meditating for over a decade now - although it's an on-off relationship. I even attended a 10-day silent meditation retreat in 2024, which has been pretty life changing for me! This is an image from the forest near the meditation hall - I walked that path probably 100 times and discovered something new every time.
I listen to lots of stuff - but Spotify knows me best, so here is my weekly discovery playlist for your perusal.
Once a month, my wife and I do a horror movie night with friends - here are the last three movies we watched (pulled from my Letterboxd account). Also, check out the recommender I built for us.
Silence 
Old Path White Clouds 
You Are Here 
No Mud, No Lotus 
Peace Is Every Step 
When Things Fall Apart 
Full Catastrophe Living 
Wherever You Go... 
Why We Meditate 
The Art of Living 
Be Here Now 
After the Ecstasy, the Laundry 
The Attention Revolution 
The Book: On the Taboo 
The Wisdom of Insecurity 
Autobiography of a Yogi 
Doors of Perception 
The Psychedelic Explorer's Guide 
Food of the Gods 
Red Rising series 
The Kingkiller Chronicle 
A Song of Ice and Fire series 
The Expanse series 
The Belgariad series 
Mistborn series 
The 13th Paladin series 
The Three-Body Problem 
Dune 
It 
Island 
Malice 
Of Blood and Fire 
Gods of the Wyrdwood 
Magician 
The Dwarves 
Uldart: Die dunkle Zeit 
A Wizard of Earthsea 
Ender's Game 
Braiding Sweetgrass 
Wild 
Into the Wild 
Into Thin Air 
Seven Years in Tibet 
Born to Run 
A Search in Secret India 
Adult Children of Emotionally Immature Parents 
How to Do the Work 
Why Has Nobody Told Me This Before? 
Why Zebras Don't Get Ulcers 
Maybe You Should Talk to Someone 
In the Realm of Hungry Ghosts 
How to Change Your Mind 
DARE 
Feeling Good 
A Liberated Mind 
Four Thousand Weeks 
Deep Work 
The Myth of Normal 
How Not to Die 
How Not to Age 
Food Rules 
Omnivore's Dilemma 
Deep Medicine 
Total Immersion 
24/6 
Smartphone Nation 
It Doesn't Have to Be Crazy at Work 
The Age of Surveillance Capitalism 
How to Break Up with Your Phone 
Make Time 
We Should All Be Feminists 
Men Explain Things to Me 
The Road to Reality 
Surely You're Joking, Mr. Feynman! 
Tools of Titans 
How to Win Friends and Influence People 
The Life-Changing Magic of Tidying Up 
When Breath Becomes Air 
Man's Search for Meaning 
That Good Night 
The Obstacle Is the Way 
Stolen Focus 
The Wim Hof Method 
Atomic Habits 
Indistractable 
Waking Up 
Built to Move 
Kitchen Confidential 
The Comfort Crisis 
The Shallows A passage from Art and Fear and what it says about how we see other people's work versus our own.
A month with Claude Max taught me how work, science, and life will change radically in the very near future.
Happy to hear from friends, collaborators, future coworkers, or people who want to talk about brains, tools, climbing, books, or dinner. Email is best.
Happy to chat about research, potential collaborations, or opportunities. Email is best. Also on LinkedIn, GitHub, Hugging Face, and Google Scholar.
