Interviews with Xiaoshi Dai, Di Xia, Feifan Gao

I currently work as a Product Designer designing AI-native financial products that help people make better long-term financial decisions. My work focuses on combining human-centered design with emerging AI capabilities to simplify complex workflows and improve user confidence.

Alongside my professional work, I also collaborate on independent research and design projects with multidisciplinary teams. These collaborations allow us to explore how artificial intelligence can create meaningful social impact beyond commercial products.

For me, design is the discipline of turning complexity into clarity. It is not only about creating beautiful interfaces but about helping people accomplish meaningful goals with confidence.

As artificial intelligence becomes part of everyday life, design plays an increasingly important role in shaping how humans understand, trust, and collaborate with intelligent systems. Great design should make advanced technology feel approachable, transparent, and empowering.

I am most passionate about designing AI-native digital products that solve meaningful human problems. I enjoy projects where technology, psychology, and business strategy intersect, especially products that help people make important life decisions.

Whether it is career development, financial planning, or creative workflows, I enjoy designing systems that simplify complexity while preserving users' sense of control and confidence.

One of my favorite projects is Sprounix AI Career Agent, an AI-native platform designed to support people throughout their career journey. Instead of treating job searching as a sequence of isolated tasks, we reimagined it as an intelligent, collaborative process where AI acts as a personalized career partner.

The project combines conversational AI, personalized recommendations, and structured career planning into a unified experience. It is especially meaningful to me because it addresses a challenge that millions of people experience while exploring new career opportunities.

Artificial intelligence is currently the technology that excites me the most. Rather than viewing AI as a replacement for human creativity, I see it as a powerful collaborator that can amplify human capabilities.

I also enjoy designing across modern digital platforms, combining design systems, prototyping, user research, and AI technologies to create products that are both intelligent and easy to use.

I feel most creative after deeply understanding people's problems. Creativity rarely comes from waiting for inspiration; it comes from observing behaviors, asking thoughtful questions, and discovering patterns that others may overlook.

Many of my strongest ideas emerge after conversations with users, sketching multiple possibilities, and iterating with teammates from different disciplines.

I pay particular attention to user trust, clarity, and decision-making. Every design decision should reduce cognitive load while helping users feel informed and confident.

Especially in AI products, transparency is essential. Users should always understand why the system makes recommendations and remain in control of important decisions.

Designing is both intellectually challenging and deeply rewarding. I enjoy the process of exploring uncertainty, testing ideas, and gradually transforming ambiguity into solutions that genuinely help people.

The most satisfying moments are when I see users accomplish something they once thought was difficult and realize that thoughtful design has made their experience simpler and more empowering.

Seeing a design come to life is one of the most rewarding moments of my work. The excitement is not simply about launching a product—it comes from seeing people use it in ways that genuinely improve their lives. When users accomplish something more confidently, make better decisions, or discover opportunities they might have otherwise missed, I feel that the design has fulfilled its purpose.

Design is ultimately about people, and the greatest satisfaction comes from creating experiences that continue delivering value long after they are released.

A successful design is measured by its impact rather than its appearance. It should solve a meaningful problem, create value for users, and support business objectives at the same time.

In the age of AI, successful design also builds trust. Users should understand how intelligent systems work, feel comfortable interacting with them, and remain in control of important decisions. When technology becomes intuitive instead of intimidating, design has succeeded.

I first ask whether the design solves the right problem. Beautiful visuals are valuable, but they cannot compensate for an experience that fails to meet users' real needs.

I also look at clarity, usability, accessibility, and long-term scalability. Good design should remain understandable as products evolve, while bad design often creates unnecessary complexity that users must overcome.

Designers have a responsibility to shape technology in ways that benefit society. As AI becomes increasingly integrated into everyday life, designers play an essential role in ensuring that these systems are transparent, ethical, inclusive, and centered around human needs.

Beyond creating products, designers influence how people learn, communicate, work, and make decisions. Every design choice carries social consequences, making responsibility an inseparable part of the design process.

Design is evolving from creating interfaces to designing intelligent systems and human-AI collaboration. Designers are no longer only responsible for screens; we increasingly shape behaviors, decision-making processes, and interactions between people and intelligent technologies.

I believe the future belongs to AI-native design, where designers work alongside artificial intelligence to create experiences that are more adaptive, personalized, and context-aware while maintaining transparency and human agency.

My most recent exhibition was at the NYCxDESIGN Festival 2026 as part of the "Becoming" exhibition in New York City, where our award-winning project Sprounix AI Career Agent was presented to an international audience. It was a meaningful opportunity to discuss how AI can reshape the future of career development through human-centered design.

Looking ahead, I hope to participate in more international exhibitions that explore the relationship between artificial intelligence, design, and society, while engaging with designers and researchers from different cultural backgrounds.

My inspiration comes primarily from observing people rather than following trends. I enjoy talking with users, understanding their frustrations, and identifying the moments where technology could better support them.

I also draw inspiration from psychology, behavioral economics, emerging AI research, architecture, industrial design, and everyday experiences. Creativity grows when diverse perspectives intersect, so I constantly seek opportunities to learn beyond the traditional boundaries of product design.

I would describe my design style as human-centered, AI-native, and systems-oriented. I enjoy simplifying complex technologies into experiences that feel approachable, trustworthy, and intuitive.

My approach begins with understanding people rather than technology. I combine user research, systems thinking, business strategy, and rapid experimentation to create solutions that are not only elegant but also practical and scalable. I believe successful design emerges when empathy, innovation, and measurable impact work together.

One of the most surprising discoveries from our research was not simply that people preferred conversation over forms, but that conversation fundamentally changed what they chose to share. Traditional application forms encourage users to think in terms of keywords and predefined categories, whereas natural dialogue allows people to explain motivations, learning experiences, transferable skills, and career aspirations that would otherwise remain invisible.This insight prompted us to redesign the onboarding experience around an adaptive conversation rather than a sequence of static fields. Instead of asking users to complete lengthy forms upfront, Sprounix gradually builds a comprehensive career profile through contextual dialogue. Each answer influences the next question, allowing the AI to explore areas where additional detail could improve understanding while avoiding repetitive or unnecessary prompts. The experience feels less like data entry and more like speaking with a knowledgeable career coach.However, conversational freedom also introduces challenges. Enterprise recruitment systems require structured, standardized information that can be compared consistently across candidates. To bridge this gap, we designed a translation layer that converts natural conversations into structured professional attributes without requiring users to think in database terminology. Rather than forcing candidates to describe themselves using recruiter-centric language, the system interprets their experiences and organizes the information into meaningful competencies, interests, and career preferences.Maintaining emotional comfort throughout this process was equally important. Looking for a new job is often accompanied by uncertainty, self-doubt, and anxiety. We deliberately designed the conversation to remain supportive rather than interrogative. The AI acknowledges achievements, asks follow-up questions naturally, and gives users opportunities to refine or correct its understanding at any point. Users are always able to review, edit, or remove generated information before it is used for matching.Ultimately, we wanted conversations to accomplish two goals simultaneously: helping people express themselves authentically while generating reliable, structured information for intelligent matching. When users no longer have to translate themselves into rigid application forms, they can focus on communicating who they are, and the technology can focus on understanding their true potential.

Traditional applicant tracking systems are highly efficient at identifying exact keyword matches, but they often overlook people whose experiences do not fit standardized descriptions. We wanted Sprounix to recognize not only what candidates have done, but also what they are capable of doing. That required moving beyond keyword matching toward understanding intent, context, transferable skills, and long-term career aspirations.From a design perspective, this meant building multiple layers of understanding rather than relying on a single resume or profile. The system considers information gathered through conversations, resumes, portfolios, projects, stated interests, and career goals, allowing AI to construct a more holistic representation of each individual. Instead of interpreting isolated words such as "designer" or "product manager," the system evaluates relationships between experiences, motivations, technical skills, collaboration patterns, and future ambitions. This richer context enables more meaningful recommendations than traditional keyword filtering alone.Equally important was designing for fairness. Early-career professionals frequently possess strong potential but relatively limited work histories, making them particularly vulnerable to conventional screening methods that prioritize years of experience or exact terminology. By emphasizing demonstrated capabilities, learning ability, project experience, and transferable competencies, we aimed to create opportunities for candidates whose potential might otherwise remain invisible.We also believed that transparency was essential for responsible AI. Rather than presenting recommendations as unexplained algorithmic decisions, Sprounix is designed to communicate why opportunities are suggested and which aspects of a user's profile contributed to the match. Giving users visibility into the reasoning process helps them improve their profiles while reinforcing trust in the system.Ultimately, our goal was not to replace human judgment but to create a more complete understanding of human potential. We believe AI should reduce bias by expanding what can be recognized, helping employers discover talent beyond conventional signals while giving candidates a fairer opportunity to demonstrate who they are rather than simply how well they optimize a resume for keywords.

One of the biggest design challenges was helping users feel that they remained in control even while their AI agent was acting autonomously on their behalf. We quickly realized that transparency could not be achieved through technical explanations alone. Simply showing that an AI had completed an action was not enough—users wanted to understand why the action happened, what information had been used, and whether they could intervene at any moment.This realization shaped the Co-pilot dashboard into something closer to a collaborative workspace than a traditional control panel. Rather than presenting automation as a hidden background process, the interface continuously communicates the AI's intentions, current activities, completed actions, and recommended next steps. Every meaningful action is accompanied by concise, human-readable explanations that answer three essential questions: What happened? Why did it happen? What can I do next? This approach transforms AI from an invisible algorithm into a visible collaborator whose reasoning users can easily follow.During iterative usability testing, we observed an interesting behavioral pattern. Participants were generally comfortable allowing the AI to perform repetitive administrative work, such as identifying suitable opportunities or preparing application materials. However, they consistently wanted explicit approval before decisions that could influence their professional identity, such as submitting applications, modifying career preferences, or communicating with employers. This led us to establish clear boundaries between AI recommendations and human decisions, ensuring that users always retained authority over high-impact actions.Another important discovery was that transparency also required uncertainty. Instead of presenting recommendations with excessive confidence, the system acknowledges when additional information would improve a match or when multiple options are equally reasonable. Showing confidence levels, assumptions, and alternative paths helped users perceive the AI as a thoughtful assistant rather than an infallible decision-maker. Interestingly, this honesty increased trust rather than reducing it.Ultimately, we believe that successful AI products should not strive to make automation invisible. They should make intelligence understandable. By designing interactions that continuously explain, invite feedback, and preserve human agency, the Co-pilot dashboard allows users to experience autonomy and control simultaneously—a balance that we believe will become increasingly important as AI systems take on more complex responsibilities in people's everyday lives.

Receiving the Iron A' Design Award was meaningful because it validated the core philosophy behind Sprounix rather than simply recognizing its visual execution. Throughout the project, we made a series of design decisions that prioritized long-term user value over short-term novelty. We focused on building an AI-native experience that balanced innovation with usability, transparency, accessibility, and human agency—qualities that we believe are essential for the next generation of intelligent products.What made this recognition particularly encouraging was the breadth of the evaluation criteria. Sprounix was never intended to be a technology demonstration or a conceptual experiment. From the beginning, we approached it as a holistic product that needed to solve a real problem while remaining intuitive, inclusive, and trustworthy. The award reinforced our belief that successful digital products should integrate interaction design, service design, AI capabilities, accessibility, and emotional experience into a coherent whole rather than optimizing a single dimension in isolation.The recognition also strengthened our confidence in pursuing ambitious interaction models that challenge conventional assumptions. Choosing an Agent-to-Agent architecture, conversational profiling, and transparent AI collaboration required us to rethink many established patterns found in traditional recruitment platforms. These decisions introduced additional design complexity, yet the evaluation suggested that innovation becomes most meaningful when it improves people's experiences instead of adding technological sophistication for its own sake.Beyond recognizing the project itself, the award strengthened our conviction that AI-native experiences should be evaluated not only by technical sophistication but also by their ability to expand human capability. As intelligent systems become increasingly embedded in everyday life, designers have a growing responsibility to ensure that innovation remains understandable, trustworthy, and ultimately empowering. This recognition encourages us to continue exploring how thoughtful design can shape a more human future for AI.

One moment from our user research fundamentally changed how we thought about the role of AI in the job search process. During an interview, a participant explained that the most exhausting part of looking for a job was not writing resumes or submitting applications—it was the emotional burden of repeatedly feeling ignored. After investing hours tailoring applications, candidates often received no response at all. Over time, many began questioning not only their qualifications but also their confidence and motivation. We realized we were not simply designing a productivity tool; we were designing for one of the most emotionally vulnerable periods in a person's professional life.This insight shifted our priorities. Instead of asking how AI could help users apply to more jobs, we began asking how AI could help them feel more supported throughout the journey. We introduced interaction patterns that emphasized encouragement, progress, and continuous guidance rather than endless task completion. Instead of presenting the experience as a checklist of applications, Sprounix helps users understand where they are in their career journey, what they have already accomplished, and what meaningful next step they can take. Small moments of positive reinforcement were intentionally designed to reduce uncertainty and maintain momentum.Gamification also evolved because of this research. Rather than relying on competitive mechanics such as points or rankings, we focused on celebrating learning, consistency, and personal growth. Progress reflects the user's own development instead of comparison with others, reinforcing the idea that career building is an individual journey rather than a competition.Ultimately, this research reminded us that technology should reduce emotional friction as much as operational friction. By combining intelligent automation with empathy, transparency, and positive feedback, Sprounix aims to make career development feel less overwhelming and more empowering. Our goal is not simply to help people find their next opportunity more efficiently, but to help them move through the process with greater confidence, resilience, and optimism about their future.

Designing a multimodal experience was never about adding more input methods—it was about allowing people to communicate in the way that felt most natural to them. During our research, we found that candidates varied greatly in how they preferred to present themselves. Some expressed their strengths more confidently through conversation, others preferred uploading a carefully prepared résumé or portfolio, while some felt more comfortable demonstrating their personality through video. We realized that asking everyone to communicate in the same way would inevitably limit their ability to present their full potential.Our design challenge therefore became one of orchestration rather than feature integration. Regardless of whether users uploaded a PDF, participated in a voice conversation, recorded a short introduction, or completed a live interview, every interaction followed a consistent structure and visual language. Instead of feeling like separate tools, each modality contributed to a unified career profile that continuously evolved as users interacted with the platform. This consistency reduced cognitive load because users always understood how each new activity contributed to improving future recommendations.Video presented a particularly sensitive design problem. Many users expressed anxiety about being evaluated on camera, especially during the early stages of their job search. Rather than making video the primary interaction, we treated it as an optional opportunity to add richer context. Before recording, users receive clear guidance about what to expect, why the recording is useful, and how it will be used. They can pause, review, rerecord, or skip the experience entirely without being penalized. Giving people this level of control significantly reduced performance anxiety while increasing their willingness to participate.Accessibility also influenced every design decision. Voice interactions include real-time transcription, uploaded documents can be reviewed and edited, and information collected from different sources is presented in a consistent, easy-to-understand format. Users never need to repeat the same information across multiple channels because the system intelligently connects insights gathered from each interaction.Ultimately, our goal was to let technology adapt to people instead of forcing people to adapt to technology. By treating every modality as another way for users to express their experiences, ambitions, and personality, Sprounix creates a richer understanding of each individual while making the overall experience feel supportive, flexible, and significantly less stressful than traditional recruitment platforms.

One of the greatest strengths of the Sprounix project was that each member of our team approached the same challenge from a different perspective. Rather than dividing work into isolated responsibilities, we treated every major design decision as a collaborative conversation where research, product strategy, interaction design, and technical feasibility continuously informed one another. This diversity of perspectives allowed us to question our own assumptions and arrive at solutions that none of us would likely have developed independently.My own focus naturally centered on user experience strategy, AI interaction design, and translating complex technologies into intuitive workflows. Other team members brought complementary strengths in visual communication, product thinking, and implementation considerations. Because everyone participated throughout the design process—from research synthesis to concept generation and iterative critique—we were able to evaluate ideas from multiple dimensions before committing to a direction.One memorable discussion involved the level of autonomy the AI agent should have. Some early concepts leaned toward maximizing automation, allowing the system to perform as many career-related tasks as possible with minimal user intervention. Others argued that users needed greater visibility and control, particularly when decisions could influence their professional identity. Rather than choosing one philosophy over the other, we developed a collaborative interaction model in which the AI proactively performs repetitive work while keeping users informed, requesting approval for high-impact decisions, and clearly explaining its reasoning. This balance between automation and human agency ultimately became one of the defining characteristics of Sprounix.These constructive disagreements were never obstacles—they were catalysts for better design. We encouraged open critique, rapid prototyping, and evidence-based decision making instead of relying on hierarchy or individual preference. Every iteration was evaluated against user needs and long-term product vision rather than personal opinions. This collaborative process not only strengthened the final product but also reinforced our shared belief that meaningful innovation emerges when diverse expertise converges around a common purpose.

Accessibility was never treated as a checklist to satisfy compliance requirements. From the beginning, we viewed accessibility as a design philosophy that improves the experience for everyone. Many of the challenges faced by people with disabilities—information overload, complex navigation, unclear system feedback, or inconsistent interactions—also affect users experiencing stress, fatigue, language barriers, or unfamiliarity with digital recruitment platforms. Designing for accessibility therefore resulted in a product that is more understandable and usable for all users.This perspective influenced countless interaction decisions throughout Sprounix. We prioritized clear visual hierarchy, consistent navigation patterns, keyboard accessibility, sufficient color contrast, descriptive feedback, and flexible content presentation. Information is communicated through multiple channels rather than relying solely on color or visual cues. Users can review AI-generated content before taking action, revisit previous decisions, and navigate complex workflows without feeling overwhelmed. These seemingly small design choices collectively create a more inclusive and confidence-inspiring experience.Accessibility also became increasingly important as conversational AI entered the product. Different users naturally prefer different ways of interacting with technology. Some communicate more comfortably through voice, others through text, while others rely on uploaded documents or portfolios. By supporting multiple input methods and allowing users to choose the interaction style that best fits their needs, we reduce unnecessary barriers instead of expecting everyone to adapt to a single interface.Looking ahead, I believe accessibility will become even more important as AI career agents become widely adopted. Intelligent systems should not only automate tasks but also expand opportunities for people who have historically faced obstacles in education, employment, or digital participation. AI has the potential to personalize experiences, simplify complex processes, and provide guidance tailored to individual circumstances. However, these benefits can only be realized if accessibility remains a fundamental design principle rather than an afterthought. For us, accessibility is ultimately about ensuring that technological progress creates broader participation, greater fairness, and more meaningful opportunities for everyone.

Looking back at Sprounix, I believe we have only begun to explore what an AI-native career companion can become. Today's career platforms still focus primarily on helping people find their next job, but careers are much more than a series of applications. They involve continuous learning, personal growth, changing life priorities, and long-term decision-making. We envision AI playing a much broader role throughout this journey—not as a replacement for human ambition, but as a trusted partner that helps people navigate increasingly complex professional lives.In the next generation of Sprounix, we hope the platform can evolve from a job-search assistant into a lifelong career advocate. Rather than becoming active only when users begin looking for a new position, the AI could continuously help identify emerging skill gaps, recommend learning opportunities, monitor industry trends, prepare users for future career transitions, and proactively suggest new possibilities based on changing goals and market conditions. Instead of reacting to career changes, the system could help users prepare for them long before they become necessary.Another exciting direction is strengthening collaboration between humans and AI rather than maximizing automation. As intelligent systems become more capable, the question is no longer whether AI can make decisions, but how those decisions should be shared with people. We believe the future lies in collaborative intelligence, where AI handles repetitive analysis while humans continue to provide judgment, values, creativity, and personal aspirations. Successful AI products should amplify human potential instead of replacing human agency.More broadly, I hope projects like Sprounix contribute to a larger shift in how society thinks about career development. Technology should not simply make recruitment faster; it should make opportunities more accessible, decisions more informed, and professional growth more equitable. If AI can reduce uncertainty, remove unnecessary barriers, and give more people the confidence to pursue meaningful work, then it has the potential to become not only a productivity tool but also a positive force for lifelong human development.