Interviews with Meiqi Zhao

My work is currently developed through an independent, interdisciplinary design practice focused on AI-powered experiences, digital products, and emotionally resonant creative systems. Rather than limiting the studio to one medium, I see it as a space for exploring how technology, interaction, and visual design can come together to create products that feel intuitive, expressive, and human.

The practice combines product strategy, user research, UX/UI design, prototyping, and visual storytelling. Many of my projects begin with a human problem or an emotional need, then evolve into digital products, AI tools, interactive systems, or speculative design concepts.

I am especially interested in designing for moments where technology becomes more personal: how people express themselves, make decisions, create, learn, or connect with others. My design studio is built around that question.

For me, design is the process of turning complexity into something people can understand, use, and emotionally connect with.

Good design is not only about form or function. It is about judgment: knowing what to simplify, what to emphasize, what to remove, and what kind of experience a person should have. A well-designed product should feel clear, intentional, and almost inevitable.

I also see design as a bridge between imagination and reality. It takes abstract ideas, cultural observations, technical possibilities, and human needs, then transforms them into something tangible. In that sense, design is both a practical discipline and a poetic one.

I enjoy designing works that combine technology, emotion, and everyday use. In particular, I am drawn to AI-powered products, creative tools, interactive experiences, and systems that help people express themselves more freely.

I like projects where the design challenge is not only visual, but also behavioral and strategic: understanding what users really need, how they move through an experience, what motivates them, and how the product can create a sense of ease, delight, or discovery.

I am also interested in designs that have a strong emotional layer. Whether it is an AI creativity product, a food discovery experience, or a digital interface, I like designing things that feel personal rather than purely functional.

One of my favorite designs is NomNow, an AI-powered food discovery product I co-founded and designed. The product was created to help people discover restaurants and food experiences through a more natural, multimodal interface using voice, image, and text.

What I liked most about this project was that it addressed a very ordinary but emotionally rich problem: deciding what to eat. Food discovery is not just about efficiency. It is connected to mood, culture, memory, social context, and personal preference. I wanted the product to feel less like a search engine and more like a thoughtful companion.

For NomNow, I worked across user research, product strategy, interaction flows, prototyping, and visual design. We interviewed more than 60 users, launched the product, and reached over 5,000 users in the first 30 days. The project also received international recognition, including the MUSE Gold Award and iF Design Award. For me, it represents the kind of design I care about most: useful, technically forward-looking, and emotionally human.

One of the first major things I designed for a company was a digital product experience that required translating business goals and user needs into a clear interface. Early in my career, I worked on product design and product management projects where I created user flows, wireframes, high-fidelity prototypes, and product requirement documents.

That experience taught me that design in a company is very different from design in isolation. It is not only about having a strong concept, but also about communicating clearly with engineers, aligning with business priorities, understanding constraints, and making decisions that can actually be shipped.

It was an important turning point because I began to understand design as a collaborative discipline. A good designer does not only create visuals; they help teams make better decisions.

My favorite technology right now is artificial intelligence, especially multimodal AI. I am interested in how AI can work with text, voice, image, and context to create more natural forms of interaction.

As a designer, I see AI not only as a tool for automation, but as a new design material. Like light, space, or typography, AI has its own behavior, limitations, and expressive possibilities. It can make products more adaptive, personal, and creative, but it also requires careful design judgment to make the experience trustworthy and understandable.

I am especially interested in using AI to support creativity, self-expression, discovery, and decision-making. The most exciting part is not replacing human creativity, but expanding what people can make and how they interact with technology.

I feel most creative when I am moving between different fields and noticing unexpected connections. This often happens when I am reading, visiting exhibitions, observing everyday behavior, using new technologies, or having conversations with people from different backgrounds.

Creativity usually comes to me when I am not forcing an immediate answer. I like to collect fragments first: a visual reference, a user pain point, a technical possibility, a cultural pattern, or a personal emotion. Later, these fragments start to connect and form a clearer direction.

I also feel very creative during the early stage of a project, when the problem is still open and the possibilities are not yet fixed. That uncertainty gives me space to explore, question assumptions, and imagine new forms of experience.

During the design process, I focus most on clarity, emotional resonance, and the relationship between user behavior and product structure.

Clarity is always the foundation. A design should help people understand what they can do, where they are, and why the experience matters. I pay close attention to information architecture, interaction flow, hierarchy, and the removal of unnecessary friction.

At the same time, I care about the emotional layer of design. A product may be functional, but if it feels cold, confusing, or forgettable, it will not create a lasting connection with people. I try to design experiences that feel intuitive, personal, and meaningful.

I also focus on whether the design can be realized and improved over time. Good design should not only look strong as a concept; it should work in real situations, support business goals, and continue to evolve through user feedback.

When I design, I often feel a mix of curiosity, focus, tension, and excitement.

The early stage usually begins with curiosity. I want to understand the problem, the users, the context, and the hidden patterns behind people’s behavior. As the project develops, the process becomes more focused and sometimes intense, because design requires many decisions and trade-offs.

There is also a productive kind of uncertainty. Before an idea becomes clear, I often need to stay with ambiguity for a while. I see that as an important part of the creative process.

When the design direction finally starts to make sense, I feel a quiet sense of momentum. It is the feeling that separate fragments are beginning to form a coherent experience.

When my designs are realized, I usually feel a combination of relief, responsibility, and motivation.

There is relief because an idea has moved from an abstract concept into something real. At the same time, realization also makes the design more accountable. Once people start using it, the work is no longer only about my intention; it is about whether the design truly serves users in real situations.

I also feel motivated because launch is not the end of design. It is the beginning of learning. Seeing how people interact with the work often reveals things that could not be fully predicted during the design process.

For me, the most meaningful moment is when a design quietly becomes useful in someone’s life. That is when I feel the value of the work most clearly.

A successful design solves a real problem in a way that feels simple, meaningful, and appropriate to its context.

It should meet user needs, but it should also have a clear point of view. I do not think successful design is only measured by aesthetics or usability alone. It is the combination of usefulness, clarity, emotional quality, technical feasibility, and cultural relevance.

A successful design should also create trust. People should feel that the product understands them, respects their time, and behaves in a predictable way. This is especially important in AI-powered products, where users need transparency and confidence in the system.

Ultimately, I think a design is successful when it becomes almost invisible in use, but still leaves a clear impression emotionally or intellectually.

When I judge a design, I first consider whether it understands the real problem. A visually polished design can still be weak if it is solving the wrong problem or ignoring the user’s actual context.

After that, I look at clarity and structure. I ask whether the design makes the experience easier to understand, whether the hierarchy is logical, and whether each design decision has a clear purpose.

I also consider emotional tone. Good design has a certain sensitivity to how people feel while using it. It should not only be efficient; it should feel appropriate, respectful, and memorable.

Finally, I consider execution. Details such as typography, layout, interaction, motion, and consistency matter because they reveal the quality of the thinking behind the work.

Designers have the responsibility to consider the consequences of what they create. Design shapes behavior, attention, consumption, and the way people relate to technology, culture, and each other.

For society, designers should create work that is accessible, ethical, and respectful of different users. This means avoiding unnecessary manipulation, reducing confusion, and making systems easier for people to understand and control. In AI-related design, this responsibility becomes even more important because designers influence how people trust, use, and depend on intelligent systems.

For the environment, designers should be aware of material choices, production methods, product lifespan, and digital sustainability. Even digital products have environmental costs through infrastructure, energy use, and consumption patterns.

I believe designers should not treat beauty and responsibility as separate goals. The best design should improve experience while also reducing harm.

The design field is moving from designing static objects and interfaces toward designing intelligent, adaptive, and system-level experiences.

In the past, many design problems were focused on screens, products, or visual communication. These are still important, but the field is expanding. Designers now need to understand data, AI, behavior, ethics, business models, and complex ecosystems. The role of a designer is becoming more strategic and interdisciplinary.

I believe the future of design will be deeply shaped by artificial intelligence. AI will change how designers research, prototype, generate ideas, and build products. However, this does not make human design judgment less important. It makes it more important. As tools become more powerful, designers will need stronger taste, clearer ethics, and a deeper understanding of human needs.

The future of design is not only about making things look better or work faster. It is about designing more humane relationships between people, technology, and the world around them.

The central design challenge was to make advanced AI feel invisible without making it feel simplistic. We wanted users to benefit from machine learning, multimodal recognition, and adaptive estimation, but we did not want the product to feel like a technical demo.Our approach was to design the experience around the user’s daily routine, not around the model’s capabilities. Instead of asking, “How do we show everything the AI can do?”, we asked, “What is the smallest interaction that helps someone log a meal with confidence?” That shifted the design from feature exhibition to behavioral simplicity.The conversational flow became important here. A traditional tracker often makes users do the work upfront: search for food, choose from long lists, estimate portions, and manually adjust nutrition data. In NomNow, the AI carries more of that burden. The user can start with whatever input is easiest in that moment: a photo, a spoken description, or a short text entry. The system then structures the information, asks for clarification when needed, and turns a messy real-life meal into a usable nutrition record.We also tried to preserve lightness. Health products can easily become judgmental or overly clinical. For a daily-use product, the tone has to feel supportive, not punitive. The interface needed to be clear, calm, and conversational, so users could return to it without feeling that tracking was another source of pressure.For me, the sophistication of the product is not in showing complexity. It is in absorbing complexity on behalf of the user.

I believe NomNow’s use of Retrieval-Augmented Generation and Large Language Models contributes to preventive health by lowering the cost of participation. Many people understand that food awareness matters, but they do not maintain calorie tracking because the process is too manual, repetitive, and mentally demanding.RAG and LLMs help change that relationship. Retrieval can ground food information in more reliable nutritional references, while language models can interpret the messy, human way people describe meals. Together, they allow the product to move beyond a static database search and become a more adaptive assistant.This matters because preventive health often fails not from lack of intention, but from friction. A user may not want to weigh every ingredient, search through multiple entries, or correct a database every day. If the system can understand a meal from a photo, a casual description, and a short clarification, then tracking becomes much easier to sustain.I also think this approach makes health technology more accessible. People have different levels of nutrition knowledge, different cultural food backgrounds, and different relationships with dieting or wellness. A conversational AI system can meet users closer to where they are, instead of forcing everyone into the same rigid tracking model.The broader shift is from health tracking as self-discipline to health tracking as supported awareness. The product does not need to shame users or demand perfection. It can help them notice patterns, make small adjustments, and build a more sustainable relationship with food over time. That is where I see the real social value of AI in preventive health.

Winning the Bronze A’ Design Award in the Digital Product Design category is meaningful because it validates the core intention behind the Nomnow AI Calorie Tracker: to make nutrition tracking more accessible, intuitive, and genuinely usable in everyday life.From the beginning, we did not want to build another calorie tracker with a more advanced technical layer. We wanted to rethink the interaction model itself. Traditional tracking often places too much burden on users: they need to search manually, estimate portions, correct entries, and repeat the same process every day. Our goal was to reduce that burden through a conversational, multimodal experience that allowed people to log food through image, voice, and text in a more natural way.The award recognition reflects not only the product’s interface innovation, but also its user-centered thinking. It confirms that accessibility and emotional ease can be as important as technical sophistication. For a health product, the real design challenge is not simply accuracy; it is whether users can sustain the behavior over time.This validation has shaped how I think about the next phase of development. It encourages us to continue focusing on trust, personalization, and long-term habit formation. Instead of adding features for the sake of complexity, the next phase should deepen the product’s ability to understand users’ routines, food cultures, goals, and constraints, while keeping the experience light enough to become part of daily life.

One important turning point was deciding not to make the AI’s output feel overly definitive. In an early version of the product, it was tempting to let the system recognize a meal and immediately present a polished calorie result. That would have made the AI look impressive, but it would not always have been the most honest or useful experience.Food recognition is inherently uncertain. A photo may show what a dish looks like, but not the exact ingredients, portion size, cooking method, oil content, or hidden sauces. If the product pretended to know everything from one image, it could damage user trust.So we made a design decision to treat AI as a collaborative assistant rather than an unquestionable authority. The system could provide an initial estimate, but it also needed to invite clarification in a lightweight way. A user could correct the portion, add missing ingredients, or describe the meal in their own words. This made the experience less flashy, but more human and more reliable.That process taught me that thoughtful AI design is not about maximizing automation at every step. Sometimes the best design is to preserve a small amount of human input in the right place. The goal is not to remove the user from the process entirely, but to reduce the tedious parts while keeping them in control of meaningful decisions.For me, this revealed a larger principle: AI capability becomes valuable only when it is shaped by trust, context, and humility.

The origin of Nomnow AI Calorie Tracker came from my own frustration with traditional calorie logging. I had tried using nutrition-tracking apps, but the experience often felt repetitive, rigid, and mentally tiring. The hardest part was not understanding that nutrition mattered; it was maintaining the habit when every meal required manual search, selection, correction, and estimation.That firsthand experience shaped the early design language of the product. I wanted the interface to feel less like a database and more like a conversation. Instead of building the information architecture around food categories and nutrition tables first, we built it around real user behavior: taking a photo, describing a meal casually, asking questions, correcting details, and moving on.This also shaped the interaction model. The product needed to accept imperfect, incomplete input because that is how people behave in daily life. Users may not know the exact name of a dish, the weight of an ingredient, or the calorie count of a meal. The system had to support approximation, clarification, and gradual refinement.The product began to feel distinct when the experience shifted from “logging data” to “being understood.” That was the key difference. Existing calorie trackers often ask users to adapt to the structure of the app. Nomnow aimed to adapt to the way users naturally communicate about food.At that point, the product was no longer just a calorie tracker with AI features. It became a more conversational and forgiving way to build food awareness.

Over the next few years, I see Nomnow AI Calorie Tracker evolving from a meal-logging tool into a more continuous and personalized health companion.The conversational interface will remain central because it lowers the barrier to interaction. People are more likely to maintain a habit when they can engage with a product naturally, without feeling that every action requires effort. In the future, I imagine the product becoming better at understanding context: a user’s routine, dietary preferences, health goals, cultural food habits, and emotional relationship with eating.Personalization will also become more important. Calorie tracking should not be the same experience for everyone. Some users may want weight management, some may want muscle gain, some may need better awareness of eating patterns, and others may simply want to understand their meals without judgment. The product should adapt to those different needs without becoming overwhelming.Wearable integration is a natural next step. Nutrition does not exist in isolation. It connects with activity, sleep, heart rate, stress, and daily rhythm. By connecting food data with wearable signals, Nomnow could help users understand patterns more holistically, such as how meals relate to energy, recovery, or long-term health behavior.The broader potential is to make preventive care more practical. Many health tools ask users to become highly disciplined before they can benefit. I think the future of Nomnow is to do the opposite: meet users where they are, reduce friction, and turn small daily actions into useful health awareness over time.

What I hope receives wider recognition is not only the technical novelty of the Nomnow AI Calorie Tracker, but the design shift behind it: moving health tracking from a rigid act of self-monitoring toward a more conversational form of self-awareness.Many wellness products assume that users fail because they lack discipline. My experience suggested something different. Often, people already care about their health, but the tools demand too much effort, too much precision, and too much emotional tolerance. Traditional calorie tracking can make users feel judged, interrupted, or exhausted. The design opportunity was to reduce that burden without reducing the value of the insight.Through Nomnow, I became more convinced that preventive health design should not only focus on motivation, but also on friction. If a product can make a healthy behavior easier, more forgiving, and more integrated into daily life, it has a better chance of creating long-term impact.I also think the project points to an important responsibility for AI design. AI should not simply make products more powerful; it should make them more humane. In this case, the value of AI is not that it replaces human judgment, but that it helps users start, continue, and understand their own behavior with less cognitive load.Within the design community, I hope more attention is given to this kind of invisible design work: the careful choices that make technology feel less intimidating, health behavior feel less punitive, and everyday self-care feel more sustainable.

From an interaction design perspective, the main challenge was to make the AI feel like it was listening, not interrogating. Food tracking is already personal. If the product asks too many questions, it becomes annoying. If it asks too few, it may feel inaccurate or careless. The design had to find the right rhythm between guidance and freedom.Replacing the traditional search-and-scroll database model meant changing the user’s mental model. Instead of asking users to adapt to a rigid database structure, we allowed them to begin with whatever felt easiest: a photo, a short description, a voice note, or a correction. The system then interpreted, organized, and clarified the meal in a conversational way.One important design principle was progressive clarification. The product should not demand perfect information at the beginning. It should first accept the user’s natural input, then ask only for the most important missing detail, such as portion size, ingredients, or preparation method. This made the process feel lighter and more respectful.What surprised me from early users was that people responded positively not only to convenience, but also to being understood. They appreciated that they did not have to know the exact database term for a food or manually break down every ingredient from the start. The conversational dynamic made the product feel more forgiving.That feedback confirmed that in AI product design, accuracy matters, but perceived understanding matters too. Users are more willing to build a habit when the system feels adaptive to them, rather than forcing them to behave like a machine.

The guiding principle was to design around real human effort, not around technical demonstration. An AI product may impress users once with a powerful feature, but people only return when the product reduces friction in a meaningful and repeatable way.For Nomnow AI Calorie Tracker, the visual language needed to feel calm, clear, and approachable. Health products can easily become visually overloaded with charts, numbers, warnings, and goals. We wanted the interface to support awareness without creating pressure. That meant using hierarchy carefully, keeping the flow lightweight, and making the AI interaction feel like part of the daily routine rather than a separate technical process.The feedback mechanism was also important. Instead of presenting AI results as final answers, the product needed to make correction feel normal. If users adjusted a portion size or clarified an ingredient, that should not feel like the AI failed. It should feel like a natural collaboration between the user and the system.For multimodal input, the principle was flexibility. Different situations require different inputs. A user may take a photo when eating out, speak when their hands are occupied, or type when they want privacy. The product should not force one “correct” mode of interaction. It should support the easiest path in each context.What I would want an aspiring digital product designer to understand is that the difference between an impressive AI product and a useful one often comes down to restraint. You do not need to expose every capability. You need to decide what the user should feel, what they should not have to think about, and where human control still matters.Good AI design is not about making the machine look intelligent. It is about making the user feel more capable.