Playable worlds / AI learning / production systems

Playable worlds. AI learning products. Operating systems for ambitious builds.

Vincent Lin builds cinematic game prototypes, accessible AI education tools, and the behind-the-scenes systems that make complex creative work repeatable.

Live cinematic browser game

Jianghu Ascendant

A martial-arts deckbuilder you can play now.

Choose a route through a mythic wuxia world, build a martial style, and survive duels against rivals, assassins, captains, and spirits. Jianghu Ascendant is the clearest live proof of the portfolio: a playable world with atmosphere, progression, combat, and a public build.

  • Playable live build
  • Route-based roguelike structure
  • Martial-arts deckbuilding combat
  • Cinematic enemy and level art

Browser game / deckbuilder / wuxia roguelike

Project Sanctuary Logo

Narrative card battler

Project Sanctuary / cardlyB

A character-driven card battler about survival, style, and hard choices.

Project Sanctuary pairs roguelike card combat with an anime-styled party world. It shows a second path through the same portfolio promise: visual direction, systems design, playable mechanics, and production planning moving together.

  • Party roguelike card combat
  • Character-first visual identity
  • Playable prototype direction
  • Design, art, audio, and QA planning

Godot 4 .NET / C#

AI learning product

AcePrepLabs

AI practice that feels clear, supportive, and student-centered.

AcePrepLabs turns practice, feedback, and question workflows into a more approachable learning experience. The product direction is not "AI for its own sake"; it is AI that helps students know what to practice next and gives educators a clearer path to support them.

  • Student-friendly practice flow
  • AI-supported tutoring and feedback
  • Question-bank workflow proof
  • Built for families, educators, and training buyers

Behind the scenes

A production operating system for AI-assisted building.

The harness work turns ambitious projects into repeatable production loops: agents can explore, build, test, critique, and preserve context while the project keeps moving. It is the invisible layer behind faster game production, stronger prototypes, and clearer handoffs.

Coordinate

Break large work into clear roles, files, and responsibilities.

Verify

Use tests, browser checks, playtest loops, and review passes before claiming work is done.

Remember

Preserve decisions, evidence, and handoff context so complex projects do not reset every session.

Multi-role agent workflow experiments

Start with the playable work. Then inspect the proof.

The fastest way to understand the portfolio is to play Jianghu Ascendant, then follow the public project trail behind the games, learning products, and production systems.