Project 1
Landing Page
Motivation
It is this very webpage you find yourself on right now! After years of writing code just for myself, I wanted to share some of my favorite projects. And it is nice to see my work condensed in one place. So this website is up, now the only thing left to do is finishing all those other projects.
Technique
Some of my favorite HTML tools are templating engines, because of how easy they make it to populate the HTML file with structurally repetitive content, like let's say, a list of projects. Here, I used EJS for the first time. The dynamic elements are facilitated by jQuery. Everything else is pretty much good old CSS. Most of the graphic assets I found either on unsplash or iconmonstr.com.
Progress
From:
April 2022
To:
December 2022
Completion:
Tools
-
CSS
-
EJS
-
jQuery
Project 2
Trolling Trolls
Background
In November 2021, I took part in the techlabs Digital Shapers program. This is a 3 month programing course where beginners work in teams to finish a self-chosen project. Teams consist of data science, UX and web dev students. The latter was my role.
Vision
Our project Trolling Trolls was intended to be a tool for content moderators. A user is able to type in a term, sentence or hash tag and will receive a dashboard with information about the context that term is used in on the web. This should help to identfy hatespeech.
Status
Our final product has a searchbar that accepts a term. Then, tweets containing this term are collected. An AI model on the backend categorizes the tweets into different categories, before a summary is presented to the user.
Progress
From:
November 2021
To:
January 2022
Completion:
Tools
-
Bootstrap
-
Teamwork
-
Chart.js
-
MongoDB
Project 3
AI art
Background
AI has reached the average internet user. Within less then half a year, it became completely normal to use AI for writing birthday poems or producing photorealistic images of unicorns. However, using the tools to their fullest potential is in itself a skill. Much like googleing, prompting an AI tool to produce a desired result is hard. I investigated how difficult it would be to get images that could be used as video game assets.
Motivation
I came in with clear goals in mind: The AI should create pixel art images in a coherent style. I was especially curious whether I could get animation phases of a character sprite. And of course I wanted to understand what it takes to create the perfect prompts. You find the results on a designated website I created to showcase my efforts.
Conclusion
Some of the image generating AIs produce fantastic results. But if you have a very clear vision for a certain image, you should expect to spend some time trying different prompts. However, this feels like just the beginning of an AI revolution. I intensely worked with these tools for two month. In that short time span alone, the quality of some of the tools had improved notably (, and so did my prompting skills). I am excited to see how we will use AI in 10, 5 or 2 years!
Progress
From:
March 2023
To:
Mai 2023
Completion:
Tools
-
Prompt engineering
-
Vite
-
React