An AI-powered task analysis pipeline, designed to identify and aggregate skills in user-submitted tasks.

This tool hosts a Telegram bot which listens for messages describing tasks which the user has completed. On receiving these messages, they are fed into an AI model which will identify skills demonstrated.

The current form of this repo is very much a tool which I have built for myself, and not particularly user friendly. As such, the Github repo is private for the time being. In time, I will make this a deployable docker image which users deploy on their own machines.

Jotter is a terminal-based task manager. It lets you organise work into projects, each of which holds a list of tasks. Tasks can be nested under one another, so you can break larger pieces of work down into sub-tasks.

This project came about as I found myself continuously losing track of ideas for new things to work on and ideas for improvements to other projects. My thought process was 'Jira makes this really easy at work, how can I get the benefits of that, without the bloat that an individual hobbiest would not need?'. The result was this, a fairly simple Python project which will be fairly easy to add to later down the line.

My departures board project was created as a Christmas present for my dad. I wrote a simple React frontend which would go off and grab data from the Rail Data Marketplace (a really cool resource), and present this in a departure board-esque display. I loaded this onto a Raspberry Pi Zero 2b, fired it up, and found that rendering web apps on this hardware was unrealistic. After a quick vibe coded migration to Pygame, we were sorted.

This site has been designed to focus on two things which are easily overlooked during development (or at least which I find easy to overlook during development): 1 - efficiency, and 2 - simplicity. You can find a writedown of the decisions I made in creating this site on my blog.