Food for Thought 1 🍦: What do Harvard courses, product design, and BerryLine froyo have in common?

While revisiting Behaviorism, Cognitivism, and Constructivism (Ertmer & Newby, 2013) to explore ways of redesigning the program for working professionals, I was struck by how these classic learning theories map directly onto how we design products:

🍓Behaviorism: “Behaviorism focuses on the importance of the consequences of those performances”. In other words, reinforcement matters. Clear signals, feedback loops, and measurable outcomes drive adoption and habits. Think of onboarding flows as a series of small rewards that guide intentional user actions.

🍓Cognitivism: In 1950, psychologists began to de-emphasize behavioral models and shifted toward understanding learners’ conceptualizations of how learned information is organized. In other words, structure matters. People need information organized in ways that make sense. That’s why we invest in clarity in the user experience so users can focus on truly rationalizing how the product works and building lasting mental connections with the product.

🍓Constructivism: Different from behavioral and cognitive theory’s objectivistic point of view, constructivism argues that learners actively construct knowledge through experience, reflection, and interaction. Context matters. True understanding comes from hands-on experience. Real value is created when users experience and explore the product. Features should enable exploration and personalization not just scripted paths. That’s where deeper learning, and long-term loyalty, happen.

For a lighter take, as a BerryLine fan in Boston (their original frozen yogurt with mochi is the best!!), I like to think froyo captures these three theories in one cup!

🍓Behaviorism is that instant reward when the first spoonful of BerryLine makes you want more.

🍓Cognitivism is knowing which toppings and flavors fit together because you’ve learned the patterns.

🍓Constructivism is experimenting until you find your own perfect mix. 

Combining these theories, products, or even something as simple as froyo, don’t just deliver functionality. They teach. They reinforce behaviors and empower discovery. As product designers, we don’t just create products; we’re creating learning experiences for our users.

📍Food for Thought #2: Designing for Inclusion at Scale

I approached the redesign of the course for working professionals through a guiding design principle rooted in human–computer interaction: accessibility as a foundation for equity. Every design choice aimed to make learning human-centered for all learners, including those with visual, auditory, motor, and cognitive differences.

✨ Accessibility by Design (aligned with WCAG 2.1 AA, ADA, and Section 508):

❗Perceivable: Added captions, transcripts, and high-contrast visuals with descriptive alt text for all media.

❗Operable: Ensured full keyboard navigation and logical tab flow.

❗Understandable: Simplified complex educational content and introduced consistent navigation patterns.

❗Robust: Verified compatibility with WAVE web accessibility evaluation tools as well as NVDA and VoiceOver using semantic HTML to ensure assistive technology support.

But accessibility doesn’t end with interfaces and it extends to how people participate, express themselves, and create meaning online. That belief also guided my earlier work at the AI for Social Good Lab, where I led a study on how visually impaired YouTubers navigate content monetization. I interviewed 12 creators across 4 continents and found innovative strategies for visibility and income, from reframing disability representation to leveraging YouTube as a credibility portfolio for offline opportunities.

Both experiences reaffirm a core belief: Inclusive design is not just about compliance. It is about creativity, empathy, and expanding what participation looks like in digital spaces.

The HCI research community has long explored how creators from marginalized communities, including LGBTQ+, Black, and transfeminine creators (Kingsley et al., 2022; Rodriguez, 2023; Harris et al., 2023; DeVito, 2022), as well as creators with disabilities (Choi et al., 2022; Borgos et al., 2019), navigate visibility, representation, and accessibility. An MIT research group I engaged with during a PhD interview had been exploring the limitations of existing web accessibility standards, particularly their tendency to offer only basic text or table alternatives for data visualizations, which restricts non-visual data exploration (Zong et al., 2022).

If you’re interested in diving deeper into this topic to design for equity and inclusion, I highly recommend checking out these incredible works and ongoing research in this space!

📍Food for Thought #3: Learning Isn’t Just Data: What Human-Centered AI Taught Me

Over the past few weeks, I’ve had a few coffee chats with learning designers across different industries and have learned how generative AI is increasingly being used to analyze learner performance and support learning design systems. I also immersed myself in both learning design and reading human-centered AI research. These chats and reading are shifting how I think about my work as a learning designer. Recent human-centered AI research work, spanning passive sensing, digital wellbeing, AI ethics, and context-aware systems, highlights something fundamental:

✨ Learning is not just cognitive. It’s emotional, contextual, and deeply human.

As a learning designer, I often think about engagement, motivation, feedback, and learner support. But human-centered AI research made me rethink how we understand learners in the first place, and what ethical responsibilities we carry when designing “intelligent” or data-driven learning systems.

Here are two key reflections that I’m taking forward:

👾 1. Interpretation matters more than data!

Research shows that AI can easily misinterpret human behavior when it’s divorced from context (Zhang et al., 2024).

For learning design, this means:

• A drop in LMS activity does not always equal disengagement.
  
  

• A spike in late-night studying may reflect stress, caregiving responsibilities, or limited access and might not be a lack of discipline.
  
  

• Behavioral analytics must be paired with empathetic, contextual understanding of learners’ lived realities.
  
  

👾 2. AI should coach, not control!

Systems like GLOSS (Choube et al., 2025) and Trucey (Duddu et al., 2025) demonstrate that AI works best when it supports reflection, not when it dictates decisions.

For learning:

• AI tutors should help learners understand their habits, not label them as “good” or “bad.”
  
  

• Reflection prompts can help learners make meaning from their own behavior — fostering metacognition.
  
  

• The goal is empowerment, not surveillance.
  
  

📍Food for Thought #4: Designing in the ZPD — Why Support Matters More Than Structure

During the Deliver module, I found myself returning repeatedly to the learning-science concepts we explored in HPL, especially the Zone of Proximal Development (ZPD) and its implications for designing learning experiences. Seeing how people interacted with our redesign of HPL for working professionals in gallery walk made the ZPD feel directly relevant: the goal is to provide the right amount of support without overwhelming or under-supporting the learner.

Feedback from the Gallery Walk made this even clearer. Some of our text was still too dense, which risks pushing learners outside their ZPD by creating unnecessary cognitive load. Our “suggested weekly flow” also came across as more rigid than intended; instead of acting as a scaffold, it felt like a mandate, something that could actually hinder learners who need flexibility. Moreover, participants also wanted us to name the pain points of working professionals more explicitly and to build in more networking opportunities.

Overall, the Deliver module reinforced that our design should keep working professionals within a productive ZPD, which should be supported but not pressured and guided but not constrained. These insights will continue to shape how we refine our Capstone.

When I first joined the Teaching and Learning Lab Practicum to redesign How People Learn for working professionals, I wanted to better connect learning theory to authentic design decisions, build confidence working within institutional platforms, and collaborate effectively across disciplines. This project stretched me in exactly those ways.

I indeed came in with four identities: visual designer, computer scientist, researcher, and aspiring learning experience designer. This project became the place where all four had to show up at once and then negotiate with each other.

Looking back on the full arc from early prototypes to the gallery walk and refinement phase, I see this project less as “building a course” and more as learning how to design a living system: one that responds to real constraints, real people, and a rapidly changing landscape of AI-mediated learning.

My visual design background still shapes how I make information clear and inviting, but I now see aesthetics as a means of supporting cognitive clarity and accessibility rather than an endpoint in itself.

My AI/CS training empowered me to explore tools like LLM-based audio generation and AI-driven case creation, but I now approach these technologies with a more critical, ethical, and learner-centered mindset.

My research grounding in HCI reminded me to treat learner feedback, expert insights, and platform limitations as meaningful data, not barriers, to guide the design forward.

Most importantly, this project taught me to view learning experience design as an ongoing dialogue: between theory and practice, between technology and humanity, and between what learners say they need and what they actually experience. It’s a practice of continuous listening, refining, and aligning design decisions with the lived realities of the people we serve. Bringing together my backgrounds in art, computing, and research, I’m learning that great learning design isn’t about choosing one identity, and it’s about bringing them together to create experiences that truly serve learners.

Introduction and Project Overview

In the Teaching and Learning Lab (TLL) Practicum, I worked as part of a design team tasked with redesigning the How People Learn (HPL) course for working professionals. Our goal was to create a flexible, workplace-anchored learning experience that made learning science accessible, relevant, and actionable for busy adult learners.

This analysis “unpacks” the key instructional design decisions I made and how they were informed by course readings, learning theories, and the practicum’s enduring understandings:

• Learning design is all about making choices.

• If you do not design for someone, then you are designing for no one.

• Constraints are good, up to a point.

• Designers are human, designing for humans.

I focus on:

1. The audience and how it shaped my design choices.

2. The learning outcomes and their alignment with the redesigned experience.

3. The learning theories and instructional design models that guided our design process.

1. Audience: Working Professionals as the Primary Learners

The core audience for this redesign is working professionals, who are often mid-career, balancing full-time jobs, families, and other commitments. From early in the project, I treated audience definition not as a formality but as a design constraint that would anchor every decision. This aligns closely with the enduring understanding, “If you do not design for someone, then you are designing for no one.”

To deepen our understanding of this audience, I helped conduct and synthesize:

• Interviews with working professionals from both the U.S. and Asia.

• Expert interviews with learning designers from MIT and Harvard who specialize in professional and executive education.

From these conversations, several patterns emerged:

• Professionals have limited, fragmented time and often learn “on the subway,” “between meetings,” or “on their phone.”

• They want to understand “why this matters” immediately. In other words, they want to understand how each concept helps them solve real problems at work.

• Long readings and dense pages can feel overwhelming; they value micro-learning, clear structure, and quick feedback.

• They care about practical application, especially in mentoring, onboarding, decision-making, and designing learning in their own organizations.

These insights directly impacted our design choices. For example:

• We prioritized short, modular learning units rather than long, monolithic modules.

• We foregrounded workplace relevance in introductions, cases, and learning outcomes.

• We advocated for device-neutral design and pagination choices that made it easier to learn on a phone.

• We supported building low-stakes checkpoints, such as quizzes that offer only pass/fail options and appear only at the end of each module, to help professionals feel oriented and confident rather than anxious.

In short, the audience profile became a touchstone for evaluating whether our design choices were truly learner-centered instead of designer-centered.

2. Learning Outcomes and Alignment

A central principle guiding the redesign was backward design, in which learning outcomes are established first and the instructional and assessment elements are then built deliberately to support those outcomes. In our redesign of How People Learn for working professionals, I played a key role in articulating, sequencing, and aligning these outcomes to the course’s learning experiences, assessments, and media choices.

Clarifying the Learning Outcomes:

Drawing from our project charter, faculty interviews, and working professional feedback, the course’s learning outcomes were refined and sequenced to move from foundation → application → transfer. The final outcomes were:

• Outcome 1 Foundational Understanding: Learners will be able to explain key learning theories, including behaviorism, cognitivism, constructivism, and adult learning principles, in clear, workplace-relevant terms.

• Outcome 2 Applied Analysis: Learners will be able to analyze workplace learning scenarios and identify where learning theory can clarify challenges or guide decisions (e.g., mentorship, onboarding, professional development).

• Outcome 3 Reflection on Learning Practice: Learners will reflect on their own learning processes and develop a more intentional professional learning identity.

How the Redesigned Course Aligns with These Outcomes:

Outcome 1 (Foundational Understanding) Supported by Modular Concept Pages + Optional Audio + Micro-learning Structure

We intentionally designed short, modular Canvas pages that introduce each learning theory with:

• a concise definition,

• a workplace example, and

• an optional reading section of reading that connects the definition to workplace application. To increase accessibility for busy professionals, we added podcast-style audio versions for these optional readings.

This aligns directly with cognitive load theory, reducing extraneous load while ensuring foundational concepts are available in a structured, digestible form.

Outcome 2 (Applied Analysis) Supported by PingPong Chatbot–Generated Case Studies

To scaffold application, we integrated the PingPong case generator, allowing learners to enter their own professional contexts and receive tailored scenarios.

This supports Outcome 2 by enabling learners to:

• analyze a workplace-relevant problem,

• apply the week’s theory to interpret the scenario, and

• articulate reasoning grounded in course concepts.
  
  

Because cases are personalized, learners see immediate relevance, which our interviews identified as essential for busy adult learners.

Outcome 3 (Reflection on Learning Practice) Supported by Harmonize Discussions

Throughout the course, learners reflect on their own learning habits, misconceptions, and evolving understanding of how people learn. This is supported by peer-sharing in Harmonize, which also offers valuable networking opportunities for working professionals.

3. Instructional Process and Theoretical Foundations

Our design process was guided by both instructional design models and learning theories studied in the practicum:

Applying ADDIE and Backward Design

• Analysis: Through interviews, needs analysis, and collaboration on the project charter, we clarified learner needs, institutional constraints, and the scope of the redesign. I contributed to this phase by helping shape the rolling agenda and project charter that captured goals, roles, timelines, and in-scope vs. out-of-scope elements.

• Design: I helped prototype learning sequences, outcomes, and assessments, always starting with “What should learners leave able to do?” in line with backward design.

• Development: Development phase of ADDIE, I intentionally created two contrasting prototypes, one in Articulate 360 and one in a blog-style format, to test different ways of making learning theory approachable for working professionals. This dual-prototype approach allowed us to evaluate not only aesthetics and usability but also deeper pedagogical questions related to motivation, cognitive load, and learner autonomy. Working professionals told us in interviews that: they value clarity and efficiency, they do not want to read dense academic texts, they often learn in short bursts on mobile devices, and they vary significantly in how much depth they want at any given moment.
  
  Given these constraints, we needed to test two fundamentally different instructional delivery metaphors:
  
  Articulate 360 → “Interactive Micro-Learning”
  Blog Format → “Narrative Edutainment”
  
  Each prototype embodied a different design hypothesis about how adults prefer to learn theory.
  
  Prototype 1: Articulate 360 (Interactive Micro-Learning): I developed an Articulate 360 prototype that presented each learning theory through a flashcard-style micro-learning interface. Each card offered: a bite-sized definition, a quick workplace example, and an option to “expand” for deeper explanation and citations. This design tested several pedagogical principles:
  Cognitive Load Theory: Breaking content into discrete cards reduces extraneous load.
  Learner Control: Professionals can skim or dive deeper depending on available time.
  
  Feedback on the Articulate prototype revealed that professionals appreciated its clarity, modularity, and ability to scale complexity based on situational motivation. This validated micro-learning as a viable delivery mode.
  
  Prototype 2: Blog-Style Format (Narrative Edutainment)
  
  In contrast, the blog-format prototype drew from research on narrative engagement, informal learning, and the rising popularity of edutainment-style pedagogies.
  
  This prototype used a conversational tone, storytelling techniques, real-world anecdotes, and visually clean long-form text. Its pedagogical hypothesis was different: If professionals feel emotionally connected to the material and see its relevance through story, they may be more motivated to persist with content that would otherwise feel abstract.
  
  Working professionals responded positively to this prototype for different reasons: It felt human and “not academic.” Readers reported that they could imagine returning to the blog as a “resource,” not just a course page. This validated the idea that tone and narrative framing matter as much as content structure.
  
  

• Implementation: Although Canvas became the final platform, due to institutional compatibility and integration with Harmonize, the lessons from both prototypes directly shaped our Canvas implementation, the early prototypes informed our design language and content structure.
  
  

• Evaluation: We gathered feedback from working professionals, peer designers, and course faculty (including Professor Matthew L. Miller), using that feedback to iterate on the flow, media, and assessment strategy.
  
  

Learning Theories in Practice

Several theoretical frameworks explicitly shaped our decisions:

• Andragogy: We respected adults’ need for relevance and self-direction by clearly connecting each concept to workplace problems and giving them choices in how they engage (e.g., optional readings, audio versions, and flexible asynchronous work).

• Constructivism and Social Learning: Harmonize discussions and peer feedback on cases reflect the idea that learners construct knowledge socially, not just individually. Learners see how others interpret theories in different industries, strengthening transferability.

• Cognitive Load Theory and Multimedia Learning: The shift toward micro-learning, shorter Canvas pages, and a clear visual hierarchy in prototypes was driven by a desire to reduce extraneous load. In the Articulate prototype, for example, I used flashcards to separate key concepts from richer explanations, allowing learners to control the depth of their engagement.

• Situated and Authentic Learning: By centering authentic, domain-specific cases generated via the PingPong chatbot, we invited learners to work with problems that feel like their own, rather than generic examples. This supports transfer and motivation.