Encoded_meaning
Can the features of a letterform be used to encode hidden data and meaning?
Encoded_meaning is a exploration in design, computational cryptography, computer vision and information dissemination through radically offline networks.
At its core, the project allows for a seemingly innocuous message or body of text to be cryptographically encoded with hidden meaning and shared in broad daylight without fear of interception.
The project is a study in how sensitive information can be shared in a highly public forum with little concern for information leakage or interception.
This work was completed as part of my Masters thesis at NYU ITP/IMA and researched while embedded at NYU Shanghai, NYU Berlin and NYU NYC (completed 2025).
CREATIVE DIRECTION • INTERACTION DESIGN • UI/UX • RESEARCH • SOCIAL JUSTICE
How does it work?
The project stems from a theory and a provocation.
1 - Variable type is built atop latent data.
2 - What potential remains unexplored for this invisible cache of information?
Variable type allows for precise control of the visual form of a letter. A variable typeface can render a letter in any number of ways along one or more axes. The examples below are rendering along a single very expressive axis.
Building a theory and a model
Variable type allows for precise control of the visual form of a letter. A variable typeface can render a letter in any number of ways along one or more axes. The examples below are rendering along a single very expressive axis.
Crafting an interface and testing a tool
Variable type allows for precise control of the visual form of a letter. A variable typeface can render a letter in any number of ways along one or more axes. The examples below are rendering along a single very expressive axis.
Secret messages are woven into the fabric of ordinary text, each letter quietly carrying hidden meaning through imperceptible shifts in typographic weight—transforming innocuous words into a tool to carry urgent messages hidden in plain sight.
Encoded_meaning relies on a ML model to generate encoded letterforms able to be embedded able to take flight as posters, flyers, and public signage, blending seamlessly into the visual landscape of everyday life while carrying hidden truth visible to only those who know how to listen.
Characters themselves are the vehicles for data transmission. The tool uses computer vision to unlocks the invisible—scanning the subtle variations in each letterform to reveal the concealed message beneath, a bit like reading between literal lines.
The project was completed in June 2025 with the completion of my thesis and degree from NYU. With the ever-changing, lightning speed landscape of ML, AI models, and computer vision, revisiting the development of this tool could yield even greater accuracy and precision.