One of the interesting, and hard aspect of calisthenics, is the ability to progress between movements in many different ways. In contrast to classical weight training, the target is not being able to lift a certain amount of weight over a certain amount of repetitions – but being able to move into a position and hold for some time.

I was recently loooking for ways to track exercise progressions long-term. As I found existing applications not customizable enough for my needs, I instead implemented a simple interval timer. What set it aparts from others, is that executions is sourced from TOML files. Especially for intra-exercise progressions, I found a self-modifying pattern useful, where configuration keys can act on themselves.

Peeling oranges in high dimensionality $k\gg 3$ is hard. This post discusses a fundamental phenomenon in asymptotic geometry and explains at which rate you should peel an orange.

I wrote a new iteration of my note system during a week vacation from home. Motivated with two recent observations, that the system I want to built is actually a static note generator, and that Mastodon is a good fit for publishing and reviewing notes, I created a tool called ztl.

A transcription of Belle qui tiens ma vie I created with Musescore for playing on Gambe and singing together.

I recently started with some exercises of the comprehensive convex analysis and monotone operator book by Bauschke and Combettes. It is an excellent source for self-studying and a good fit for a student with EE math background.

At the beginning of the year I started practicing a new instrument and took since then classes for a bass Viola da Gamba. After around 10 month of regular sessions, I slowly developed enough dexterity to play simple tunes. This post contains a brief resume of observations of the past months and shows some music of the Gambe.

Arguably one of the most important identities in signal processing is the following theorem.

The identity is known as Moreau’s decomposition and instantiates into useful applications for variational analysis, convex optimization and signal processing. This post will explain the symbolic meaning and implication of this theorem for signal processing.

For a recent problem I had to crawl and categorize a large number of audio files from the internet. As the bottleneck was deemed to be downstream link I decided to use a pre-trained event classification model and run inference on the CPU.

Another requirement was that the crawler were running on multiple server and consistent for several weeks. My choice of language for such infrastructure heavy task is Rust and I looked into ways how to deploy an inference model.

It turned out that the process for creating a standalone binary in Rust is pretty simple. Even better, wrapping a model pre-trained on a large event classification dataset is possible. This means that in the end I could just copy the binary to any x86 machines and run them, without the need to setup virtual environments or install any packages.

In this post I will demonstrate how to convert a pre-trained model to ONNX and put it into a binary. Our goal is to create a standalone executable taking samples from stdin and classifying them into 527 known audio classes.

My favorite instructive math book I have read till now is the Cauchy-Schwarz Master Class written by Michael Steele et al.

Two features of the Tridactyl extension makes reading/writing the web much more relaxing for me. The most common element of Vim like extensions is link following with highlighted key-codes. This makes mouse pointing unecessary and reading workflow easier. (also don’t forget to use the g<f|F> chain) But I find the visualized style of many websites too distracting and use Vim in those cases to construct an answer or dissect a thread.