Projects

What I’m working on.

ikna

An open-source language-learning app using spaced repetition. It teaches phrases in context, adjusts the daily workload automatically, and stores review history locally.

I want attention to go into learning, with almost no time spent managing the card queue.

MVS Analyzer

A tool for statistical analysis of repeated measurements. It compares summary metrics through simulation on a specific dataset and estimates statistical power and false-alarm rates. It saves run settings so the analysis can be reproduced. Further development focuses on the CLI and Python API.

I first check how a variability metric behaves in the scenario I need. That makes it easier to see whether it fits the task.

Allostatic Sprint Hypothesis

I’m testing the original Allostatic Sprint hypothesis on several open datasets from attention tasks. I keep positive and null results, as well as cases where the method’s limitations prevent a conclusion.

I keep testing the model on new data and retain null results so the analysis isn’t shaped to fit the hypothesis.

Gitctap

A Python tool with no dependencies for pushing one local Git repository to several forges. It checks each remote using ordinary Git.

I made it to avoid repeating the same push by hand or setting up the same shell alias in every project.

vespian

A sleep model for Android. It tries to predict when a person will be able to fall asleep and shows a time range and confidence for each forecast. It reads Health Connect data and processes it on the device.

I’m trying to estimate when sleep is likely now, instead of setting one fixed bedtime.

trial3lib

A Compose design system for Android without Material: no elevation, ripple, rounded corners, or tonal surfaces. Shared tokens define geometry, color, and contrast, which are checked by tests.

These constraints apply across the system so I don’t have to disable Material elements by hand on each screen.