Eden Smith's UX Design Portfolio
2026 | App Design, User Experience
Khan Academy Feature Design
This feature introduces an adaptive difficulty system that adjusts in real time to each learner’s performance, helping them build confidence by meeting them at their level and providing targeted support to retain consistent learners.
I. Background & Pain Points
Students from underrepresented and low-income backgrounds often use learning platforms that assume consistent prior knowledge and high confidence. When they struggle, these systems provide limited explanation and rely on rigid progression, which can reinforce feelings of failure.
Without clear support for knowledge gaps, students may lose confidence and disengage—not due to lack of ability, but because the system isn’t built for their learning needs.
Polished for School. Eden Smith. Haiti. 2017.
II. UX Research
We conducted contextual inquiry interviews with 5 participants within our target demographic (college-aged learners from diverse educational backgrounds) to understand how they interact with existing learning platforms.
Key findings highlighted several friction points: users found the interface overwhelming, struggled with gaps in foundational knowledge, and lacked clear, actionable feedback when answering questions. These challenges often led to frustration and decreased confidence.
We synthesized insights using affinity diagramming, identifying core themes shown above: the need for a simpler UI, more inclusive functionality, improved organization, and expanded content. These findings directly informed the design direction, including the development of adaptive difficulty to better support personalized learning and confidence-building.
Affinity Diagram and Themes
III. Design Inspirations
This feature is grounded in key learning science principles that emphasize personalized, confidence-building learning experiences. Drawing from constructivism, the design encourages learners to actively engage with their understanding by making strengths and knowledge gaps visible.
The system is also informed by Vygotsky’s Zone of Proximal Development, adapting difficulty to keep tasks just beyond a learner’s current ability. By acting as a “more knowledgeable other,” the system guides students toward targeted support where it’s needed most.
Primary research further shaped the design. Findings on educational inequality highlight how limited support and gaps in foundational knowledge can impact learning outcomes. Additionally, work by McGee and Martin shows that students may internalize struggle as a reflection of identity rather than a solvable gap.
Together, these insights informed a system that reframes mistakes, reinforces missing concepts, and supports both skill development and learner confidence.
Lo-fi Sketch
Mid-fi Screens
Storyboard
skill map was removed in mid-fi iteration as it was a bit unrelated to core functionality
IV. User Testing and Evolution
Throughout the cycle of the project, my team conducted contextual inquiries to improve the feature.
i. Feedback from lo fi protoypte
The low-fidelity prototype introduced the core adaptive difficulty flow. During the post-lesson quiz, the system identified patterns in missed questions, named the skill gap, and directed users toward the lesson area that needed more support.
Feedback gathered:
-
Users wanted clearer explanations for what they got wrong.
-
Users needed the app to explicitly name the skill gap.
-
Automatically opening the related lesson felt useful, but needed more context so users understood why they were being redirected.
ii. Feedback from mid fi prototype
The mid-fidelity prototype added question-by-question feedback, visual cues for correct and incorrect answers, and clearer messaging around which skill needed improvement.
Feedback gathered:
-
Explanations were helpful, but needed stronger reasoning for why answers were correct.
-
The feature required more in-app guidance so users could understand it without verbal explanation.
-
The visual design needed to feel more polished and engaging.
iii. Feedback from hifi prototype
The high-fidelity prototype refined the experience with clearer navigation, improved visual feedback, and a more responsive adaptive difficulty system.
Feedback gathered:
-
Users found the flow easy to navigate and understand.
-
Immediate feedback after incorrect answers helped guide learning.
-
Users appreciated that the system responded to their individual performance.
-
Remaining feedback focused on adding deeper explanations and further improving the UI.
Final Prototype
Deisgn Feedback from Final Testing and Presentation
-
consistent spacing
-
using smart animate for smoother transitions
-
keeping consistent font sizes
-
re-considering color choices (greens)
Reflection
What went well
My group worked very well together. We always made sure to communicate to meet deadlines on time. I enjoyed the user testing sessions and they helped me pick up new things that I wasn't thinking about before. I was also new to the user centered design process and this was a great intro to that. I believe I'll take this experience with me to heart as a new designer. All the assignments were necessary for iteration and improvement of the design.
What went ok
I wish that my group had taken a stronger focus on aesthetic design earlier in the semester, I think that would have created a stronger, more cohesive project in the end. Overall I think we could have made sure to make things more standardized instead of just working on our features by ourselves.
What we needed more time on
We needed more time during the final stretch which was the aesthetic and polish portion.
What I would change if I were designing the assignment
I would've introduced aesthetics on the first Figma prototype. It would have made it less time consuming, towards the end. My group was under the impression that we had more creative control over the look of the application. Instead of taking so much time thinking about what things should look like and trying to make it aesthetically make sense for the assignment task we should have copied the skeleton of Khan Academy mobile application and tweaked it to fit our features. I think there was a false sense of control that we thought we had over the aspects of our design in terms of how they would translate to underrepresented groups or low-income groups.