When learning to read should feel safe, joyful, and possible

In an AI-driven learning environment, that meant designing not just for engagement—but for predictability, trust, and developmental safety at scale. Early literacy is one of the most consequential stages of learning — and also one of the most fragile. Small moments of confusion or frustration can shape how children see themselves as learners.


Dragonfly was created to support K–5 students by combining storytelling, visual warmth, and adaptive AI systems into a literacy experience that feels encouraging rather than evaluative. The challenge was not just to design a product — but to design a system that could scale creativity, maintain trust, and protect developmental needs.

At A Glance

Role: Principal Product Designer, AI Systems

Partners: Literacy Experts, Learning Science, Engineering, Media Production

Timeline: Multi-phase initiative

Audience: K–5 students and educators

Focus: AI-generated media systems, literacy scaffolding, ethical design


System Scope & Scale

  • AI-generated visual narratives used across K–5 literacy experiences

  • Reusable character and environment system supporting hundreds of story variations

  • Designed for multi-year reuse across grade bands and international markets

The Real Challenge

Scaling imagination without losing intention

The core challenge wasn’t generation—it was control.
Dragonfly required an AI system that could scale content creation without introducing visual inconsistency, cognitive overload, cultural bias, or instructional drift for early readers.


Key Challenges

  • Creating age-appropriate, emotionally safe AI-generated media

  • Maintaining visual and narrative consistency at scale

  • Supporting multiple literacy levels without stigma

  • Ensuring AI outputs aligned with instructional intent and DEI standards

My Role & Scope

System Ownership & Design Leadership

I served as the design owner for Dragonfly’s AI-generated media system—defining how characters, environments, and visual narratives were generated, constrained, and scaled across the product.


Direct ownership

  • Owned the end-to-end AI media system design

  • Defined generation constraints, guardrails, and reusable visual frameworks

  • Established prompt and environment standards used across teams


Strategic influence

  • Set ethical AI guidelines for young learners

  • Created visual language standards adopted across multiple teams

  • Balanced creative flexibility with instructional clarity and safety


Collaboration

  • Worked with literacy experts to align AI output with pedagogical goals

  • Partnered with engineers on scalable AI workflows

  • Enabled international teams through documentation and systemized assets


This role functioned as a Staff-level design owner, setting durable standards rather than producing one-off solutions.

Impactful Decisions

Where restraint mattered more than possibility

Character consistency

Prevented visual drift that would have required ongoing manual QA across hundreds of assets


Cognitive load limits

Reduced learner distraction and supported comprehension for early readers


Ethical guardrails

Minimized bias and ensured age-appropriate representation at scale


Production constraints

Enabled predictable output across teams, reducing iteration cycles and rework


These decisions reduced long-term production risk and allowed teams to scale content without increasing review complexity or cognitive burden on learners.

How Dragonfly Shows Up for Learners

Supporting confidence, not performance anxiety

Because AI-generated content can feel unpredictable to early readers, visual consistency was treated as a cognitive safety requirement—not a branding preference.


Visual storytelling was designed to:

  • Reinforce comprehension

  • Support independent learning

  • Reduce intimidation around reading

  • Create moments of delight without distraction

Outcomes & Signals

What this system unlocked

Dragonfly enabled teams to:

  • Produce consistent, high-quality media at scale

  • Maintain ethical and instructional alignment

  • Adapt content across grades and regions

  • Reduce manual asset production time

  • Create a cohesive visual identity students could trust


Educators reported that students felt more confident and engaged — particularly those who typically struggled with reading.

Designing AI with care

Dragonfly reinforced my belief that AI design is not about capability — it’s about responsibility. Especially in education, systems must be designed with empathy, intention, and restraint. This project continues to shape how I approach AI as a design material. This project shaped my approach to AI design as a practice of systems thinking, ethical restraint, and long-term stewardship—not novelty.