I'm a biologist and bioinformatician who kept following the data: from the lab bench to Python scripts,
from research servers to an OpenStack cloud, and now into systems programming at 42 Heilbronn.
What I want to do next is help organisations move their processes onto cloud infrastructure they can trust.
I started out in the lab.
Before I wrote software, I learned how science stays trustworthy: precise protocols,
careful documentation and results that someone else can reproduce. That is still how I work.
B.Sc. Biology
University of Tübingen
Biology taught me to think in systems: cells, pathways and organisms, where small changes ripple through everything else.
As a research assistant in cell culture I worked to strict standard operating procedures.
Every passage, medium change and deviation was documented, because one undocumented step can make
weeks of work unreproducible.
That habit carries over directly to validated IT systems in regulated industries such as pharma, where proving
a system does exactly what it should matters as much as the system itself.
M.Sc. Bioinformatics
University of Tübingen
Modern biology produces more data than any lab notebook can hold. In bioinformatics, biological questions
become computational ones, and the answers depend as much on the pipeline and the infrastructure as on the biology.
This is where I shifted from using computers for science to building the systems
science runs on. My thesis at the Center for Quantitative Biology (QBiC) ended up being an infrastructure project.
Automating lab analysis
Python, research assistant role
Recurring measurement series were analysed by hand in spreadsheets: slow, and easy to get subtly wrong.
I replaced that with Python scripts that run the same analysis the same way every time.
This was a small project, but it was my first real piece of digital transformation: a manual process, made faster,
less error-prone and reproducible by putting the logic in code.
Then the data needed somewhere to live.
For my master's thesis I went one layer down: away from the analysis and into the infrastructure underneath it.
OpenStack cloud for research data
Master's thesis, Cloud Computing Engineer at QBiC Tübingen
Research groups needed to exchange sensitive data securely between servers. I designed and operated an
OpenStack-based private cloud (infrastructure as a service) to make that possible:
Keystone for identity and access, so only authorised users and services reach the data
Nova to provision compute instances on demand instead of hand-configuring servers
Neutron for segmented virtual networks, keeping traffic between projects isolated
Cinder and Swift for block and object storage
Encrypted transfer between servers, and continuous monitoring and tuning so it stayed stable under constant load
OpenStack
Linux
Virtual networking
IAM
Monitoring
Where I want to go: cloud and transformation.
Building that cloud showed me what interests me most: not one tool, but the moment
an organisation changes how it works because its infrastructure finally lets it.
Cloud infrastructure
IaaS, identity and access management, network segmentation, storage. I've run these in practice,
and I want to go deeper into hybrid and public cloud platforms and infrastructure as code.
From paper to platform
Labs and factories still run on spreadsheets, paper and isolated systems. I want to help map those processes,
find the gaps and move them onto systems that scale.
Regulated by design
In life sciences, a system must be secure and also provably correct. I'm currently reading into GxP, GAMP 5
and computer system validation, so compliance is part of the design from the start.
Now I build software from first principles.
Since January 2025 I've been at 42 Heilbronn. There are no lectures and no teachers: you teach yourself each topic,
build the project, and then other students review your code line by line before it passes.
It's the most effective way of learning I've found.
C and the Piscine
Rebuilding the C standard library, memory and pointers, Makefiles and debugging from the shell.
Unix and algorithms
Processes, pipes and signals, sorting under constraints, and hardening a Linux server.
Shells, threads, networks
A bash-like shell built as a team, concurrency with mutexes, IP addressing and subnetting.
miniRT
A ray tracer in C: geometry, lighting and shadows from first principles.
C++ Modules 00–07
OOP from the ground up: orthodox canonical form, inheritance, polymorphism, exceptions, casts, templates.
C++ 08 and Webserv
STL containers, iterators and algorithms. Webserv, a small HTTP server in C++, is next.
In progress, October 2026
miniRT: a ray tracer in C
42 Heilbronn, completed
miniRT renders 3D scenes from a plain-text scene file. For every pixel it sends a ray from the camera into the scene,
works out what the ray hits first, and calculates how brightly that point is lit.
Ray intersection maths for spheres, planes and cylinders
Ambient and diffuse lighting, with hard shadows from a second ray cast towards the light
A parser for the .rt scene format that validates every value
Vector maths written from scratch, rendered to a window with MiniLibX
It's the most maths-heavy project I've done: linear algebra and geometry turned into working, testable code.
C
Linear algebra
Ray tracing
MiniLibX
The stack I build on.
Almost all of this I learned after my studies, at 42 Heilbronn and through self-study,
working from the operating system upwards.
Web
HTML semantic, accessible markup
CSS layout and animation without frameworks, like this site
Applications and scripting
Python scripting, automation and data processing
Java object-oriented application development
Shell Bash scripting and daily command-line work
Systems programming
C libc rebuild, a bash-like shell, threads, the miniRT ray tracer