- C/C++
- Python
- JS/TS
- Java
- MatLab
- VHDL
About
As an academically trained computer engineer (B.Eng.), designer and future maker (M.A.) with a focus on AI, I'm looking for interdisciplinary connections between art, computer science and philosophy. My computer science background involves a range of skills, from bare metal C/C++ to high-level frameworks like PyTorch.

In the field of Future Making and Philosophy, I specialize in radical constructivism. This is an epistemology that holds the view that one's own reality is a construction. This may sound strange at first, but epistemology especially in the form of constructivism and computer science have an unimaginable and fascinating intersection. In addition to it, I participate in interdisciplinary art exhibitions and have a strong interest in literature.
- Machine Learning
- Natural language processing
- React
- Cellular Automata
- Conversational User Interfaces
- μC Development
- Absurdism
- Radical constructivism
- Second-order cybernetics
- Philosophy of mind
- Laws of Form
- Dadaism
Projects
The paper demonstrates how to simulate visually and intuitively the dynamics of Spencer-Browns Laws of Form using a cellular automaton. First conceived as a tool to visually simulate the Laws of Form, the attempt unfolds certain philosophical implications that could be relevant to the discourse of the form within the form. Among other things, it will be shown how one can intuitively simulate a full adder and parts of the reentry in this cellular automaton.
For the project, rules were derived for a cellular automaton from which the axioms of Spencer-Brown's system can emerge. A simulator was then developed for demonstration purposes. Konrad Zuse proposed in his paper »calculating space« what might be the internal structure of the universe, namely a cellular automaton (a discrete mathematical model). In the same year 1969, Spencer-Brown independently proposed, through his Laws of Form, a system that has the distinguished observer as its underlying and self-generating structure (see the LoF simulator).

Interestingly, the necessary rules for the automaton are far more complex than Spencer-Brown's postulated axioms. The implications should be clear.
What is »good« and what is »bad«? How should one live his life? What is the purpose? Does this or that really exist? What does it mean to talk about this or that? What does it mean to understand? Can one really understand? Where does one get the certainty that something apart from himself - including others - really exists? These are questions about one's own life, all related to the belief in an Archimedean, absolute point. The belief in possible objective knowledge, in a real reality. When Albert Camus states the principle „[...] the way of acting of an honest person must be determined by what he considers to be true [...]“ [1, p. 7], then it is clear that an honest person needs truths as objective anchors. But then the question arises, for this person, where these anchors can be found.

In the search for the possibility of absolute cognition, consequently in the search for the answer to the watzlawickian question »how real is reality?«, a variety of disciplines, ranging from neurobiology to systems theory to the philosophy of language, were traversed. However, no possibility of objective cognition was found, on the contrary, the belief was even rejected, in the spirit of Heinz von Foerster's words that objectivity is the delusion that observations can be made without an observer. Everything is and will be self-referentially constructed, the whole of one's reality, including others and the self. The only criterion: viability, something fits or it does not. All these insights are related to a dizzying term, the title of this work.

The consequences of no objective reality which is accessible to us do not lead to a numbing nihilism and a lonely apathy, but absolute freedom and responsibility, a new non-dogmatic sense of values for religion, aesthetics, art, science, and other world views. Finally, the dizzying term provides a non-dogmatic approach to the meaning of life, in the spirit of Albert Camus' eternal revolt. A new sense of value, a new self-awareness of one's constructive nature emerges.
[1] Albert Camus. Der Mythos des Sisyphos: Ein Versuch über das Absurde. Reinbek bei Hamburg: Rowohlt, 1999.
In 1969 George Spencer-Brown published the work »Laws of Form«, in this work he tried to build arithmetic, algebra, and logic based on a single operation, the distinction. This distinction, also called mark or cross, is at the same time an operator as well as an operand, which is applied in space (here it's an infinitely large white area). If there is a cross in space, i.e. a distinction has been made, then this space is marked. However, if one goes into the area that was distinguished by the distinction (inside and outside), it will be unmarked again unless you introduce another distinction there.

Every mathematical system is based on axioms, simply accepted and fixed rules. In LoF there are exactly two axioms. Making a distinction twice is equivalent to making a single distinction and a distinction inside a distinction annuls both distinctions. Each expression can be transformed into exactly one of two states, the marked or unmarked space.
LoF also offers non-negligible epistemological implications in the field of radical constructivism. It provides a solid basis for explaining autopoietic systems (Humberto Maturana and Francisco Varela), deriving theories of observation and system theory (Niklas Luhmann), and cybernetic concepts (Heinz von Foerster). Moreover, there are opinions that the statements presented in Laws of Form are the basis of all cognition.
This simulator offers several functions to get an intuitive understanding of the dynamics of Laws of Form.The art manifesto as a literary form expresses itself through non-conformity and opposition. It declares intentions and proclaims world concepts, not being satisfied with the claim of describing the world but intervening in it and realizing its demands by stepping into the world. The work deals with different modes of cultural production and examines techniques of power.

Based on stochastic calculations from a database of 500 techno-anarchist manifestos, a single, repetitive mantra emerges after 20650 iterations: "Why do you know our world. Cyberspace does not understand the state... ". This becomes the title of the work. For this project Linux-based drivers for thermal printers and self-learning methods based on neural networks and Markov chains for text generation were developed.
Epistemia was a temporary three-day socio-economic utopian space created together with Theo Gomes and Joshua Burghardt in which free thought, participation, and discourse were the three pillars of this utopia.
In the sense of radical constructivism, each person has his or her own very personal reality, constructed by him or herself. In social discourse, these realities meet and create together an intersubjective reality, which defines social norms, laws, values, truths and social interfaces. These interfaces determine at the same time which realities - which individuals - are connectable to this intersubjective reality, the discourse.

In order to demonstrate this process to the individuals and to raise their awareness of it, an experimental set-up was developed. People were guided to explore their own reality and capture it with the help of semantic maps. These semantic maps, which represented parts of their reality, could then interact with other realities on the »discubus«, a cube symbolizing the medium of discourse. As a result, an increasingly complex discourse gradually unfolded, creating both opportunities and restrictions as well as its own specific rules.
In addition to this, there were also other participatory experiments and contents on the topics of the economy of the common good, radical constructivism, discourse, and public space in the project space. For each interaction in Epistemia, people could collect points, which were then converted into charitable donations.
Flat - Round, Left - Right, Liberal - Conservative, Believer - Atheist, Rational - Irrational. Some say, there is the world conspiracy. The others say, all fantasies, it does not exist. Conspiracy or not, the world must be saved, that is a fact, say some. No problem, science will save us, say the others. Science has its limits and has created more problems than it has solved, change must come from us, retort some. Exactly, Jesus loves and will save you as long as you believe, preach the others. How can Jesus save us if we are Islamized by foreigners, some shout. All nonsense, don't stress, it's too late anyway, live who can, say the others. For every predicate there is a negation - so there is also one for every personal opinion, for every opinion and for every world view. Assuming that one's own reality also consists of one's opinion, one's personal attitude and one's own view of the world, we have to admit that with nearly eight billion individuals there can certainly be more than one reality.
We humans classify, archive, give names, try almost like by an obsessive-compulsive disorder to interpret an apparent order into the world, in which possibly there is none. Every living being, every thing, even things that are not things, are assigned acoustic and graphic identifiers, the main thing being that one could point at them and proudly make sounds. With all these designated things, which are mostly of linguistic nature, the human being is so convinced to perceive the real one reality. There may be doubts whether the thing that is designated in the everyday life is an X or a Y. It may also happen that we are not aware of it. It may also happen that we say X and others object in the opposite way by saying Y. A single perceived reality, which is of course one's own, has to remain consistent in itself, but contradictory identifiers or relations are anything but consistent. The only possibility to create consistency in these cases, to bring reality in order, is to adjust one's own opinion or to reject the other. In the latter case, also called difference of opinion, conflict is often pre-programmed, if only by holding a small grudge against the other individual. It may also happen more rarely that the signified is neither X nor Y nor any other known signifier, in these cases the human compulsive neurosis of wanting to give a name to everything, both material and non-material, comes to full effect. With radiant joy of discovery, a combination of sounds is tried out and assigned to the (un)thing, often in the hope that other individuals will adopt these sounds. However, what may also not be next to several realities is that things could remain nameless and thus cannot be put into a linguistic order.
But haven't we just admitted to ourselves that there are potentially eight billion realities, which are still to be named, mind you? Why not classify these, archive them, bring them into an order with each other and thereby recognize that one's own reality is one of many, no more or less important, that most realities are possibly unique and moreover, through the process of classifying, at the same time let one's obsessive neurosis live out more? Instead of compulsively putting one's reality in order, why not put one's reality, one's reality, in order?
This is precisely where the reality database project comes in, joining all the other countless databases, from gene databases to product databases. This digital database, which can be accessed via the Internet, allows each individual to classify his or her reality, archive it and put it in order with other realities. For the classification process semantic maps are used, for the ordering machine learning methods are used, which among other things cluster and calculate similarities between different semantic maps. This makes it possible to compare one's own reality with the reality of others, to recognize where there are compatible labels in order to find common topics of conversation, where there are labels with which one has differences of opinion, in order to enter into discourse there, or to find completely new topics and labels that expand one's own reality, which one would not otherwise have discovered oneself. Vaccination causes autism, the earth is flat, Corona is a lie? Clustering in this case would bring one's reality close to other similar realities and suggest new possible appropriate identifiers or themes for one's reality such as jet fuel can't melt steel beams or lizard people. Or maybe one is tired of one's own reality and wants a change, in which case one can find countless inspirations in the database. Or maybe you realize that you are not so unique after all, or more positively not alone, and that you largely overlap with millions of other realities.
By default anonymous, in order to contain the possible dystopian scenarios somewhat, the realities can also be stored with the personal identity for a good reason. Politicians could, for example, present their reality to the voters and the voters could start a direct comparison of which politician best represents their reality. The repeated redefinition of one's own reality over a period of decades is also possible; the ontogenesis of one's own reality would thus be traceable. To the own childhood photos perhaps one day childhood realities would gel. The phylogenetic change over decades or hundreds of realities could be analyzed more easily by historians in the future and anthropologists and ethnologists would have a new tool with which they could analyze entire cultures today.
The reality database is financed by donations and managed by a non-governmental organization. In the event of government intervention, the project is to be deleted immediately, as the risk of the project falling into the wrong hands and being misused for dystopian scenarios is too great.
Narrative Economics is the study of how narratives convey values and emotions, how they spread like pathogens, thus influencing society and the economy. In order to understand recessions, for example, it is not enough to understand the cybernetic control loops behind them, but one must also include the human factor with its irrational, emotionally driven actions. Above all, this includes our belief in stories that create meaning, the narratives. Simulating the dynamic system behavior of narratives helps us to understand the underlying dynamics of history, why people acted the way they did, and thus history itself. These infectious narratives behave similarly to pathogens and can be simulated with the help of mathematical models.

This web app was specifically developed to simulate the dynamic system behavior of narratives in society using cellular automata and mathematical modeling. By using cellular automata, in which the individual cells are able to represent character values such as political views, skepticism, liberalism, or existing knowledge through mathematical modeling, it is possible to simulate this complex system behavior in a simplified way and thus get a feeling of how narratives spread in society. Narratives can emerge spontaneously, evolve, be in competition, or symbiosis with other narratives, and can also be forgotten or dismissed. The user can design different forms of societies, such as pluralistic, populistic, or a society that is skeptical towards any form of societal change. Each color represented here represents a narrative, the more similar the colors the more similar the narrative. Thus, through the complex interaction of narratives and their unfolding, beautiful and fascinating landscapes emerge. Explore the dynamic spread and development of narratives in different social systems.
Listen to the pure information of any kind of DNA interpreted as a piece of music. This web app was developed to allow you to rise above being a determined product of DNA and be an interpreter and composer converting our biological blueprint into a piece of music.
Furthermore, instead of always reading dry words and structural formulas when dealing with biology, it would be interesting to listen to soulful sounds that have their origin in the blueprint of life. Not everybody understands structural formulas, but everybody understands music, even if this understanding is not rational, but rather sensual.

You can combine different instruments with each other, thus creating orchestral pieces. With a little sensitivity, you can even get musical experiences out of pathogens, like HIV or COVID-19. Through artistic interpretations of (one's own) DNA, man freely defines its meaning. Contrary to fixed value judgements such as good or bad, high or low, proficient or lazy, he distances himself, places himself as an interpreter and no longer as a determined, transparent product and thus as an individual above this sequence of elements. He becomes more than the sum of his bases.
This project is a reactive web app that simulates a variety of common epidemiological models. Epidemics can be modelled mathematically in order to make forecasts and, if necessary, take countermeasures in time.
The most known epidemiological model was developed in 1927 by William Ogilvy Kermack & Anderson Gray McKendrick, which divides the population into three groups: susceptible (S), infected (I) and recovered (R) and also models the interactions between them. Infected people have an infection rate β to infect other susceptible people. β models both the chance of contact between people and the chance of successful transmission. Infected persons can recover with a recovery rate γ and belong to the group of recovered persons. The recovery rate can represent both the mortality rate and the recovery rate, since in this model, the recovering individuals are only characterized by the absence of infectivity after infection, and this can be assumed from both dead and immune individuals. μ models here the normal population dynamics and describes how many people die (not related to the disease) and are born, thus, the dynamic behavior of an epidemic can be represented in the model. I₀ is the percentage of the initial population which is infected.

Besides this probably best-known model, there are other mathematical models available. This web app is mainly intended as a learning tool to get a feeling for epidemiological models.
This application can algorithmically convert brain waves into real-time art. For this purpose, algorithms in the field of signal processing were developed for a commercial electroencephalogram to process the raw data, obtain the frequency bands (alpha, beta, theta, etc.) as well as the origin of the brainwaves in real-time, and then transform them into dynamically generated art.

This project is intended to represent an antipode of the reductionist neurosciences, by using their tools not to degrade subjective aesthetic experiences to their mere brain processes, but rather, to emerge aesthetic experiences from their brain processes in a complex reciprocal interaction of feedback loops. After a period of training, most of the test subjects were able to dynamically change the art with their thoughts. A total of six different algorithmic interpretations were developed, which differ in the interpretation of the brain waves.
The cybernetic world model of Jay Wright Forrester, the founder of system dynamics, published in 1971, attempts to predict the future growth of humanity. Population, natural resources, environmental pollution, agricultural, and capital investment represent the five levels that interact through complex feedback loops in order to fully describe global human growth and quality of life. For example, the population is regulated by both birth and death rates, and both rates in turn interact directly and indirectly with all other system parameters, including themselves, through feedback loops that are interleaved with each other.

The model can be simulated and shows that our exponential growth cannot be sustained in this way and that in all scenarios a collapse of humanity in the 21st century can only be prevented by immediate action (1971). All constants were chosen so that the simulation from 1900 to 1970 corresponds to reality. The model served as the basis for the more complex world model world3, which was the object of the book The Limits of Growth (Club of Rome) published one year later. Just like the original from 1971, it is possible to run through potential scenarios and future predictions with the help of the modifiers.
The promise of artificial intelligence (AI), in particular its latest developments in deep learning, has been influencing all kinds of disciplines such as engineering, business, agriculture, and humanities. More recently it also includes disciplines that were "reserved" to humans such as art and design. While there is a strong debate going on about whether creativity is profoundly human, we want to investigate if creativity can be supported or fostered by AI.

This project created together with Marco Limm and published in two follow-up papers investigates if AI is capable of inspiring designers by suggesting unexpected design variations, learning the designer's taste or being a co-creation partner. To do so we created several AI and genetic algorithms, which can be trained by a small sample set of shapes of a given object, to propose novel shapes.
The evaluation of our proposed methods revealed that it can be used by trained designers as well as non-designers to support the design process in different phases and that it could lead to novel designs not intended/foreseen by designers. In addition, the AI-based democratic design developed here might be an important tool for target group-oriented design in the future. The following bottles were designed by AI algorithms.
Examination phases, stress, or just holidays. There are situations in which you simply cannot get to water or care for your plants. In times of IoT and control engineering, this should no longer be a concern. This project deals with that problem.

With the help of this fully working self-developed greenhouse, the soil moisture, air humidity, soil temperature, air temperature and illumination are fully automatically supervised and regulated with the help of the specially developed control technology. The only thing that is required from the user is the input of desired values like air temperature via the Human-Machine-Interface. Thanks to the persistent storage of data in the familiar CSV format, measurement series for example for biological research purposes can be easily carried out. Via a Wi-Fi connection, data can be read and set using a browser. Status notifications can be sent via email and Twitter.
Sorting is one of the most important tasks in computer science. Whenever you want to search efficiently in large amounts of data, the data should be sorted. Google organizes websites for fast access, on shopping websites you can sort results by any criteria and the whole backend of the internet with its various databases uses sorting for fast data access. There are countless sorting algorithms with specific advantages and disadvantages, all of which are objects of research in theoretical computer science.

This project allows to visualize, compare and evaluate different sorting algorithms ranging from Introsort over Cyclesort up to Quicksort in real-time and in various scenarios. With the help of freely configurable parameters, different scenarios can be easily tested. For example, the number of available threads can be varied, or the copying and comparison effort can be adjusted. All sorting algorithms have been self-developed in modern, high-performance C++ and can be used as C++ templates freely for other projects if desired.
This project is a complete retro game for the arcade console »Vectrex« released in 1982. The 8 bit console has 1 KB Ram, 4 KB ROM and 1.5 MHz clock frequency. The screen is a monochrome CRT monitor (combined with a plastic overlay for static colors), which inserts additional timing and drawing constraints due to its relative positioning of the electron beam.
The complete game including levels, assets and music must run within these specifications and is therefore written in high performance C and assembly code. Jump tables, pointer bithacks, inline assembly, overlapping arrays, duffs device and much more allow the game to provide as much content as possible within these specifications. The game itself is yet another arcade racer (Yarg).
Publications & Exhibitions
George Spencer-Brown's »Laws of Form« (LoF) provided an important formalism and important insights for all kinds of fields in 1969. He made it possible to calculate with layered, self-referential spaces, time, and even the »blank« space using the distinction as a basal operation. However, with a few exceptions in sociology (Niklas Luhmann & Dirk Baecker) and mathematics (Louis H. Kauffman), cybernetic biology (Humberto Maturana & Francisco varela) the work has found few practical use cases. User interfaces, on the other hand, be they graphical or e.g. conversational, are practically designed, described and most importantly used every day - they are the experiential interfaces of our (digital) environments. Although some formalisms for these have been developed in recent years, they are on the one hand very domain-specific and on the other hand do not allow to calculate with the interfaces and thus our (digital) environments. It turns out that, in the spirit of LoF, the fundamental operation for user interfaces is also the distinction. In this paper, a formalism for user interfaces based on LoF shall be worked out, which allows to describe, compute and (domain-spanning) transform them solely by means of the distinction. In particular, it will be shown that this is an exceedingly natural way to describe and reason about (self-referential) graphical user interfaces (which can be thought of as distinctive layered spaces) and to find fundamental connections between them and other disciplines. This work will maybe lay the foundation for future designers to later develop automatic theorem provers for interfaces and to perform formal verifications.
How did painting change with the advent of photography? There are all kinds of viable explanations for this. In this work, however, another suitable explanation is to be presented. Photography is said to have placed in front of the painters the self-referential mirror that reality itself cannot be represented even in the most realistic representations. That reality is something never attainable by the photographer. That the mechanized man, with his eyes resembling a camera obscura, can never recognize a real reality and therefore a fortiori painting. These findings, which result solely from the observation of the camera, provide a possible explanation - which itself makes no claim to truth - as to why the styles from Impressionism to Surrealism may have developed. Not because photography replaced the realistic image, but because it made the impossibility of a representation of the real apparent. In the end, perhaps it can be said that photography liberated paintings from the bondage and illusion of realism and provided new possibilities of which we did not see that we did not see.








Academics
- Overall grade 1.0 (A+, excellent)
- Thesis: »f = f(f)?«
- Overall GPA 4.0 (A+)
- Art & Design
- Exploring interdisciplinary boundaries
- Overall grade 1.1 (A+, excellent)
- Thesis: ARTificial Intelligence - Automation of the artistic design process through artificial neural networks and genetic algorithms
- Change of course to computer engineering
Work
- Build blockchain APIs both for the internal web development team and, in the future, for community development projects.
- Developing and maintaining all Node.js server-side network components in JavaScript/TypeScript.
- Designing user-facing UI and back-end services for various business processes.
- Implementing effective security protocols, data protection measures, and storage solutions for web3.
- Machine learning for natural language processing for GAIA-X MERLOT
- Conversational User Interfaces / Chatbots
- Recommender Systems based on genetic algorithms and NLP
- Full stack development
- iXperience Lab - intelligent interaction & immersive experience lab
- Introduction to the ethical, epistemological and social issues of AI and the basic concepts.
- Introduction to machine learning algorithms from SVM to Neural Networks.
- Analyzing (future) use cases of AI
- 2D signal processing and AI development for radar sensors targeting building automation and gesture recognition.
- Reinforcement learning for football robotics.
- Integration of AI simulation environments.
- Supervision of teaching labs for C++, C, MATLAB and metrology.
- Software development of IoT solutions for embedded systems (medical Smartwatch) in the field of medical technology in C.
- Development of server-side Linux software in Python.
- Supporting the data evaluation of a clinical study.
- Tutor for foundations of computer science and C-Programming.