DK●
← Back to home
DAKOTA KRUG / SELECTED WORK

Ideas in
motion.

Explore practical tools and AI coursework, with a focus on workflows, experimentation, and clear documentation.

01 / WELLNESS WEB APPPERSONAL PROJECT
Find your
stride.
Stride’s runner artwork: an athlete sprinting in a neon green shirtYOUR PACE. YOUR PATH.

Stride

01

A little progress, every day.

A personal wellness journal that brings running, food habits, mood, and body check-ins into one place. I wanted an approachable way to capture daily habits and reflect on progress without switching between separate logs. The project combines structured entries, weekly planning, and exportable records in a browser-based experience.

HTML / CSS / JavaScriptProgressive web appBrowser storage
Inside the project +

The idea — Make everyday wellbeing easier to reflect on through a single, approachable interface.

The experience — Running logs, weekly plans, goals, mood check-ins, journal entries, and data export. An animated welcome gives the app its own athletic personality.

Design consideration — Journal data stays in the visitor’s browser in the Sites version; export provides a backup. Browser-local storage does not provide account sync.

AI-assisted personal project.

Open Stride ↗Demo access is currently restricted.
02 / MUSIC & INTERACTIONEXPERIMENTAL PROTOTYPE
Keyline intro screen featuring an illustrated pianist in a red dress playing a grand piano

Keyline

02

Making music visible.

An experimental piano-learning app combining sheet music, animated keys, and illustrated hand guides. It explores how sound and visual feedback can work together to make practice easier to follow. Adjustable playback and separate-hand focus let learners break a piece into smaller steps, while an experimental transcription workflow explores the possibilities and limitations of machine learning for music.

JavaScriptBrowser audioMachine learning
Inside the project +

The idea — Connect what a learner hears with what they see on the score and keyboard.

The experience — Beginner exercises, adjustable playback, separate-hand focus, and synchronized visual playthroughs. An experimental local audio pipeline creates draft piano arrangements, with MIDI, MusicXML, and PDF exports.

Design consideration — Audio transcription and illustrated finger positions are estimates. Practice progress is browser-local and completion is self-reported; the app does not assess performance accuracy.

AI-assisted personal project.

Open Keyline ↗Demo access is currently restricted.
03 / GITHUB SPOTLIGHTTEAM JAM PROJECT · NOV 2026
Illustrated armored swordsman from Darkest Knight
Darkest
Knight.AN HONORABLE DUEL TO THE DEATH
THREE.JS / CLAUDE CODE / TEAM BUILD

GitHub

03

Latest: Darkest Knight.

A team-built browser game from the November 6, 2026 Anthropic Game Jam at SCAD. A timing-based parry mechanic drives two boss encounters, with an emphasis on attack timing and difficulty. Working within a short jam format meant iterating on a focused mechanic, testing how encounters felt, and refining the experience as a team. It is an example of my interest in directing AI-assisted development and learning through practical builds.

JavaScript / Three.jsAI-assisted developmentTeam collaboration
Inside the project +

The build — Created with Claude Code as part of a three-person jam team. The team’s reflection describes directing the art and attack system, adjusting timing and speed, and choosing a more challenging boss pattern.

The experience — Two boss fights, title and end screens, restart controls, and touch controls. The emphasis is on reading attacks and responding at the right moment.

Project status — The game was submitted to the jam gallery. Its detailed AI-use disclosure and running AI-use log remain unfinished in the source project.

Team project · AI-assisted code and artwork. This is a creative AI project alongside my career focus in AI and data analytics.

Explore my GitHub ↗Play Darkest Knight ↗
04 / APPLIED AICOURSEWORK & EXPERIMENTS
ASK / TEST / REFINEAI101.

From prompts
to practical workflows.

VISUAL ITERATION / DOCUMENTS / AGENTS

AI-101

04

Learning AI by putting it to work.

A collection of coursework exploring how generative AI can support creative experiments, document workflows, and early agent concepts. The projects move from producing and refining visual outputs to organizing learning materials and drafting multi-step workflows.

This work supports my focus on applied AI: asking clearer questions, reviewing what a tool produces, and documenting where human judgment is still needed. These are learning experiments, with iteration and review forming part of the process.

Prompt engineeringGenerative AIWorkflow design
Inside the coursework +

Visual iteration — Experimenting with image prompts and revisions to explore how direction changes an output.

Document workflows — A study-guide workflow, including WorkflowQuest, exploring how source material can be organized into a more useful learning format.

Agent experiments — Early drafts for agent-based workflows. These remain coursework prototypes with review still pending, rather than finished autonomous systems.

Individual coursework · AI-assisted experiments and documentation.

View AI-101 repository ↗The repository currently requires access.