Where NP-hard problems meet industrial-strength guesswork.

‍Here's MaphSAT: a jointly-programmed SAT solver. Here's an earlier version with cactus plots. One of my favourite courses so far.

Programming in C: come for the pointers, stay for the buffer overflows.

‍Here's an implementation of Reversi in C, completed during an extensive practical course on systems programming. The implementation included handling signals, shared memory, pipes, preventing memory leaks, etc.

Where everything works fine until n>10.

‍Here's a fascinating SAT-encoding problem I've once solved as part of a joint project.

‍Here's a paper (including a presentation and exercises) from a BSc seminar on cake-cutting protocols.

Here's a presentation from my favourite course at ETH Zürich, Models of Computation. I still think the amount of homework and grading were incommensurate, but the weekly challenges were wonderful.

Why just draw shapes if you can overthink them into oblivion?

Here's my short report on plane tilings from the Geometry: Combinatorics and Algorithms course at ETH Zürich. Here's a presentation on the same topic. The report and presentation were just a fraction of the overall course workload.

Here's my presentation on Simplicits, which was a partial requirement for the Digital Humans seminar.

Wasserstein distance: the art of paying extra to move dirt in the fanciest possible way.

Here's my report from a research course on TDA. Here's the code and my part of the presentation. The project involved graph embeddings in six dimensions, convex sets, MDS, SQLSP, and a bunch of other pompous abbreviations now blissfully banished from my memory.

Just glorified curve fitting, except now we do it with GPUs.

ML was my minor at ETH Zürich. Felt like beating Jumanji, modulo the rampaging hippos.

Here are my essays, assignments, and a presentation for the AI in Education course at ETH Zürich.

Here are projects 1 and 2 from the Introduction to ML course, and projects 1 and 2 from the Advanced ML course.

Because turning backprop into a differential equation somehow feels more pretentious.

‍Here's a poster I once did as part of a research course. I was responsible for the optimal transport section.

Tossing linguistics for layers and data and proving you can master language without understanding it.

Here's a brief report from my MSc practical project. A painfully uninspiring exercise in LLM fine-tuning.

Here's a joint report for the Computational Intelligence Lab at ETH Zürich. This was just a partial requirement, the course also ended with an exam.

‍Here are all my presentations from the Philosophy of Language and Computation I/II courses. They also required some essays, but those were rip-offs from my previous papers.

Where you spend decades debating the meaning of meaning and still can’t agree on it.

Here are my Cambridge essays and a presentation on the syntax of idioms.

Here are my two BA theses: one on formal semantics, the other on pragmatics.

Let’s debate whether numbers exist instead of actually doing anything useful with them.

Here are my essays for the Sampler of Histories and Philosophies of Mathematics course. They also required weekly readings of way too many pages and writing several comments of the texts.

Here's my essay for the Philosophy of Physics: Complex Systems course. Oh, wait, I still need to write it. Update: I somehow made it and submitted the paper with 32 seconds to spare before the deadline. Felt like Philomena Cunk while writing this essay.

A cesspool of wildly disparate assignments.

I’ve found myself on the eternal scavenger hunt for credit points, so I had to complete the following short courses, while watching F1 in the background:
‍Technology Investing, Technology and Entrepreneurship, Patenting Digital Innovations. Those courses also included a mandatory presentation.

Here is one project from the Formal Methods and Verification course.

Over the years, I’ve completed other assignments spanning hundreds of pages, but I’ve shredded those — probably for the best.