Louisville Institute of Technology (LIT) – Louisville, KY

Louisville Institute of Technology builds practical AI, automation, cybersecurity, and workforce technology learning through real implementation and human-centered systems.

Skylar Tran | Build, Learn, Improve — AI Adventure

PROJECT JOURNAL · November 15, 2026
Skylar’s ideas, design decisions, and learning through building.

BUILD, LEARN, IMPROVE

Curiosity becomes something you can play.

Skylar Tran is a student creator exploring how games can help kids learn about AI. This portfolio follows a real project: the idea, the feedback, the revisions, and the work still ahead.

This introduction and the explanations below were prepared with AI assistance. Quoted requests are Skylar’s own words from the app-building chat.

PROJECT 01 / WORKING PLAYABLE PROTOTYPE

AI Adventure

Lessons lead to quizzes, and knowledge helps a mage explore themed dungeons. Skylar proposed the learning-and-battle concept and directed its evolution through specific feedback.

🌲 Forest → 🪸 Reef → ❄️ Glacier → ☁️ Sky islands

Learn → Answer & understand → Battle → Review & advance

Play AI Adventure ↗

Play as a guest without an account. Progress saves on this device. An eligible parent or guardian can sign in for cloud saves.

The starting question

“My thoughts for this app is to teach kids how to use ai like you and grok and other ai”

Skylar then proposed quizzes, points, abilities, monsters, and dungeons. The project gained a concrete audience and a repeating activity: learn something, apply it, and keep exploring.

Who did what?

Skylar: originated the concept and name; requested learning and combat systems; evaluated visuals and layout; identified confusing behavior; asked for revisions and bug checks.

AI assistant: wrote and revised application code, expanded draft lesson and question content, implemented interactions and artwork, and performed automated checks. This is not a claim that Skylar wrote every line of code or every lesson unaided.

Six decisions that shaped the app

01 / Remove a quiz shortcut

“the right answer is the long one all the time”

Observation: answer length could reveal the correct choice. Response: revise answer choices and increase challenge. Skylar later requested explanations for wrong answers. This demonstrates attention to whether the quiz rewards understanding. Balanced wording still needs review with real learners.

02 / Connect mana recovery to learning

“add a restore mana ability that makes you do a question about that topic in that dungeon”

Skylar requested replacing potions with a topic question while limiting the equipped loadout to four abilities. Restore Mana connects a battle resource to recalling the current topic. This joins two systems around the game’s learning purpose.

03 / Make attacks communicate impact

“when you press the fire ball it just looks like your mana goes down and the opponents health goes down”

Changing numbers alone did not communicate impact. Requested improvements included visible spell effects, hurt expressions, opponents facing the mage, and nearby health and mana bars. This uses visual feedback to make the rules easier to read.

04 / Keep controls within reach

“make the abilities more higher up on the screen so you dont have to scroll all the way down”

Skylar identified excessive scrolling during battle. Abilities moved higher, and the interruption screen between defeated waves was removed. These changes address both layout and pacing.

05 / Make environments distinctive

“each level has a theme so it doesnt get repetitive. for example and underwater dungeon”

Distinct themes replaced a repeated backdrop. Skylar also critiqued the fish’s appearance, then asked for the same animal-specific treatment and animation across bosses. This shows evaluation, refinement, and applying a standard consistently.

06 / Review the whole journey

“after part five can you add the final test which makes you do a test about all the parts 25 questions”

The resulting cumulative assessment covers all five parts before the last boss. The Master Mage celebration follows the last victory. This considers the whole learning journey beyond adding another battle.

These excerpts preserve Skylar’s wording. The explanations interpret documented requests; they do not invent private thoughts or prove measured learning gains.

Inside the prototype

  1. AI Explorer: prompting, truth, privacy, and creativity.
  2. How AI Works: patterns, training, prediction, and mistakes.
  3. AI Creator: stories, images, sound, and code.
  4. AI Detective: bias, misleading output, evidence, and fairness.
  5. AI Builder: planning, prototyping, testing, and launching.

The current build contains 20 dungeons, 120 regular quiz questions, a 12-question review for each part, and a 25-question final test. These are content counts, not evidence of educational effectiveness.

Project proof card

Status: publicly playable prototype. Portfolio assembled November 15, 2026 from the development chat and current source.

Audience: kids learning AI foundations through lessons, quizzes, and a fantasy battle game.

Tools: AI coding assistant; HTML, CSS, JavaScript, SVG artwork, and a server-backed hosting setup. Skylar supplies direction and feedback; the assistant implements and checks the software.

Evidence: publicly playable build, retained source/version history, specific request excerpts, and automated regression checks.

Verification: on November 15, 2026 the assistant reran all 11 current automated regression suites; all passed. These cover progression, inventory, stun behavior, curriculum structure, environments, creator permissions, combat state, final-exam persistence, and the graduation gate.

A concrete bug: an exam redraw could reset answer feedback and score state. The assistant corrected the state handling and added a regression check. These are assistant-run software checks, not tests written or independently performed by Skylar. Passing does not mean the app is bug-free.

Limits: no measured learning improvement or independent user study is established. Cross-device sign-in has not been fully verified end to end. Further work includes completing the adult-account sign-in test, reviewing lesson accuracy and quiz wording, and testing with permitted users.

Learning journal

The requests move beyond adding features: Skylar spotted a quiz shortcut, explained unclear battle feedback, asked for controls that fit the screen, and tied resource recovery to a learning question. Each observation gave the assistant a more specific problem to solve.

Skylar’s own reflection, November 15, 2026: Asked which change he was most proud of, Skylar replied: “the first one becuase it was the base of this whole app and i worked on it the most”. The phrase “the first one” is preserved without assigning it to a specific feature. Still to add: what Skylar can explain or demonstrate without AI answering for him.

A profile for every future app

Each real project will get a proof card recording its purpose, Skylar’s contribution, tools and helpers, first attempt, changes and reasons, tests and results, permitted demo, limits, learning, next step, and parent release review. Only AI Adventure is documented here; future projects will be added when there is actual work to show.


No private school records or personal contact details are included in this project journal. This portfolio documents effort and growth; it makes no promise about university admission.

Scroll to top