Before I understood supply chains, I understood that someone was waiting for an ingredient. My father works in food trading, which brought me into contact with many restaurants. Through that connection, I began to understand how much work happens before a meal reaches the table. A restaurant depends on people coordinating deliveries, tracking inventory, communicating clearly, and keeping promises. Customers may never see that work, but they experience the consequences when it goes wrong. The lesson I will carry throughout my life is that responsibility means understanding who depends on your work. An order is more than a row in a spreadsheet. For the person receiving it, it can determine what they can serve and how smoothly their day runs. That lesson became especially important during the Wuhan lockdown in 2020. When offline business channels collapsed, I worked with company leadership to build a digital operating system within 72 hours using Excel, DingTalk, and custom scripts. We connected orders, inventory, customer communication, and settlement in a workflow people could actually use. At its peak, the system handled more than 3,000 orders per day and reduced manual errors by 95%. The experience taught me that hard work becomes more effective when people have reliable systems behind them. I carried that principle through my education at Tsinghua University and later into leading a multimodal AI team at Pika. Whether I am working with food businesses or developing an AI product, I ask the same question: whose day will become easier if we get this right? With $100 and one completely free Saturday, I would invite my father to spend the day exploring a local food market and sharing a meal. I would budget $40 for lunch, $35 for ingredients, $15 for transportation, and $10 for dessert. I would ask him about the early days of his business, the mistakes that taught him the most, and the people who helped him along the way. Then we would cook dinner together. Much of my professional life involves moving quickly and solving problems. A free Saturday would be a chance to slow down, listen, and enjoy the food that first connected me to the working world. If I could help a restaurant add one dish, I would propose a mushroom-and-tofu rice bowl with ginger, garlic, a light soy glaze, and crisp seasonal vegetables. Chili oil would come on the side so customers could choose their preferred heat. The mushrooms would provide a savory depth, the tofu would make it satisfying, and the vegetables would add freshness and texture. I would want it to be affordable, comforting, and practical for the kitchen to prepare consistently. Drawing on my food-trading background, I would also consider ingredient availability and waste before suggesting it. Food taught me to connect ambition with usefulness. In my future career, I want to build technology that respects the people doing everyday work—and gives them more time, fewer avoidable problems, and greater room to grow.

这份经历最有说服力的主线是：**从武汉疫情期间的企业数字化，到前沿 AI 产品研发，再把技术能力带回中国，服务资源有限的企业和年轻人。** 我把第三人称提名材料改成了第一人称申请文，保留核心成果，将重点放在未来贡献上。 **先确认资格：该奖学金要求申请者正在美国攻读本科。材料写的是“清华大学毕业生”，未说明你目前是否符合这一条件。** 下文也请仅在这些经历确实属于你、数字可核实时使用；没有加入未经提供的经济困难。 英文正文约 **530 词**： --- During the Wuhan lockdown in 2020, I learned that technology’s greatest value is sometimes very practical: helping a business keep operating when everything familiar stops working. As offline sales channels disappeared, the company I was working with had no engineering team, no external funding, and no digital infrastructure. Orders, inventory, customer communication, and payments needed a new system immediately. Working directly with company leadership, I built one within 72 hours using Excel, DingTalk, and custom scripts. At its peak, the system processed more than 3,000 orders a day. It reduced manual errors by 95% and shortened the order-to-cash cycle by 60%. Within six months, online channels accounted for 85% of sales. The business eventually reached ¥50 million in annual revenue and expanded to 300 employees. Those numbers mattered, but the deeper lesson was about access. A traditional business did not need a large technology department to benefit from digital tools. It needed someone who understood its problems, could build under pressure, and would take responsibility for whether the solution worked. That experience shaped how I hope to contribute to China through my career. My education at Tsinghua University gave me a foundation for approaching difficult problems systematically. My professional experience taught me to connect that thinking to implementation. Later, at Pika, I led a 12-person multimodal AI research and development team, working across technical direction, experimentation, and product decisions. We delivered five major product demos in two months, exploring combinations of text, image, and audio generation. One attracted more than 200,000 overseas views and over 4,700 shares. Working in both traditional business and frontier AI showed me a gap I want to help close: powerful technology is advancing rapidly, while many smaller organizations still struggle to use it effectively. My long-term goal is to build affordable, accessible AI products for small and medium-sized businesses in China. I want to help owners automate repetitive administrative work, understand their operations, and reach customers more effectively. For a business with limited resources, reducing errors or receiving payments sooner can create room to invest, hire, and grow. I plan to begin with specific workflows rather than broad promises. By partnering with businesses, observing their daily challenges, and testing tools with employees, I can identify where AI creates measurable value. Success would mean fewer mistakes, less time spent on repetitive tasks, and tools people can use without specialized training. My lockdown experience gives me a practical starting point; my AI experience expands what I can build. I also want to share this knowledge with students and aspiring entrepreneurs beyond major technology hubs. Through practical workshops and mentorship, I hope to teach them how to identify a real problem, build a modest first solution, and improve it through feedback. These skills can help young people create opportunities in their own communities. I do not measure my ambition only by the sophistication of the technology I develop. I measure it by who gains the ability to do more because of it. My goal is to bring the knowledge I gain back to China in forms people can use—helping businesses become more resilient and helping young people become builders themselves.

At the beginning of the 2020 Wuhan lockdown, I faced a problem I had no idea how to solve. An offline trading business I was involved with suddenly had to operate almost entirely online. Orders were still coming in, but the systems people depended on—paper records, phone calls, spreadsheets scattered across computers—were no longer workable. There was no engineering team to call, no software budget, and no time to wait. So I started tinkering. I spent the next 72 hours turning whatever tools I could access into a makeshift operating system. I connected Excel spreadsheets, built lightweight scripts, automated repetitive tasks, and used DingTalk to create workflows for order tracking, inventory, customer communication, and settlement. Whenever I encountered something the system could not do, I tried to find a workaround, build a small tool, or connect two things that were never designed to work together. The most satisfying moment was not when the system was finished. It was watching other people use something that had not existed three days earlier. Employees who had been manually checking orders could suddenly see what needed to be done. Managers could track operations remotely. Customers could continue placing and following orders without relying on a physical office. What had initially looked like an impossible disruption became a system people could actually depend on. Eventually, the platform processed more than 3,000 orders a day, reduced manual errors by 95%, and shortened the order-to-cash cycle by 60%. Within six months, 85% of the company’s revenue was coming through online channels. But what stayed with me was the process, not the numbers. I realized that I loved the hacker mindset: when a tool does not exist, build one; when the obvious approach fails, try something strange; when two systems do not fit together, figure out how to make them talk. Computer science gave me a language for this instinct and a way to turn curiosity into something other people could use. I carried that spirit to Tsinghua University, where I studied computer science and worked on multimodal AI. Later, at Pika, I led a 12-person multimodal AI team and rapidly experimented with new generative video products. Today, I work on foundation models and world models, exploring how AI can understand and interact with the physical world. Kyle’s story resonates with me because technology is most exciting when it makes us ask, “What if we tried this?” That is the kind of engineer I want to become: curious enough to explore what others overlook, stubborn enough to build what does not yet exist, and generous enough to make what I build useful to other people.

My interest in computer science began not in a classroom, but during the 2020 Wuhan lockdown. When the city shut down, an offline trading business I was involved with suddenly had to operate almost entirely online. There was no engineering team, no software infrastructure, and no time to wait for someone else to build it. I started with what I had—Excel, DingTalk, and simple scripts—and built a digital operating system in 72 hours. It eventually processed more than 3,000 orders a day, reduced manual errors by 95%, shortened the order-to-cash cycle by 60%, and helped move 85% of the company’s revenue online within six months. More importantly, I discovered something that has shaped my life ever since: computer science is not simply about writing code. It is a way to turn complexity into systems, and systems into real-world impact. That realization led me to pursue computer science at Tsinghua University, where I focused on artificial intelligence and multimodal learning. Since then, I have worked on generative AI and audio-visual intelligence, led a 12-person multimodal AI engineering team at Pika, and contributed to research recognized as a CVPR Highlight. I have also continued exploring how AI can understand the physical world through world models and foundation models. My long-term goal is to become a technical entrepreneur building AI systems that move beyond the screen and meaningfully interact with the real world. I want to combine deep computer science research with the ability to build products and organizations that reach millions of people. I believe I would be a strong recipient of this scholarship because computer science has never been an abstract academic subject to me. It gave me a way to solve problems when existing systems failed, and it changed the direction of my life. I hope to keep using that ability to build systems that create opportunities for others—just as the opportunity to study computer science created one for me.

I did not come to the United States because I had a clear path waiting for me. I came because I had learned how to build a path when none existed. I grew up in a complicated family and entered university without the kind of roadmap that makes an ambitious career feel inevitable. My family initially believed computer science was too difficult for me and pushed me toward another field. At Tsinghua University, when the research opportunities I wanted were not readily available, I learned to create them myself. I taught myself unfamiliar technical areas, formulated my own research questions, reached out to researchers, and built opportunities independently. That persistence transformed my education. I graduated from Tsinghua with a GPA approaching 3.9/4.0, published three first-author papers, received major academic honors including a Presidential Scholarship, and conducted research at MIT. I was later admitted to graduate programs at Harvard and Yale. These experiences showed me that education can fundamentally expand the set of futures available to a person. My path has also extended beyond academic research. During the 2020 Wuhan lockdown, I built a digital operating system for a traditional trading operation in 72 hours, with no engineering team or external funding. The system eventually processed more than 3,000 orders per day, reduced manual errors by 95%, shortened the order-to-cash cycle by 60%, and helped move 85% of revenue online within six months. Later, at Pika, I led a 12-person multimodal AI R&D team developing generative video technology. Within two months, we delivered five major product releases, including one that reached more than 200,000 overseas views and 4,700 shares. These experiences shaped why I want to continue my education in the United States. I want to study at the frontier of AI, particularly multimodal and world models that can help machines understand physical and social environments. The United States offers an unusually concentrated ecosystem of researchers, universities, entrepreneurs, and technologies at this frontier. I want to learn from that ecosystem while also contributing my own perspective and experience. As an international student, I also see education as a bridge rather than a one-way journey. My experience moving between Chinese and American academic environments has shown me how much can emerge when different technical traditions, cultures, and perspectives interact. In the long term, I hope to build AI research and companies that connect talent and ideas across countries rather than keeping innovation inside a single ecosystem. I have also tried to make the opportunities I received available to others. In 2022, I volunteered with Tsinghua University's Program Buddy Group, supporting underrepresented students interested in coding. I know what it feels like to enter a technical field without knowing whether you belong. I want to continue mentoring students and helping people from less-connected backgrounds access technical education and research opportunities. This scholarship would help reduce the financial barrier to continuing my education in the United States and give me greater freedom to invest in research and technical development. Education changed my trajectory by giving me the ability to build a future that others had not imagined for me. I hope to use that education to build technologies, connect people across borders, and create opportunities for the next generation.

Growing up, I learned that a family can love you and still be a difficult place to grow up. My parents separated when I was young, but their relationship did not simply disappear. They continued living together for many years, and I often found myself caught between them, trying to understand two different versions of the same family. I became highly sensitive to conflict and unusually careful about communication. Over time, these experiences affected how I trusted people, expressed myself, and understood my place within a family. As I grew older, these difficulties became intertwined with my own mental health. I have struggled with bipolar disorder and with the challenges of communicating and functioning while dealing with it. There were periods when maintaining stability required enormous effort that other people could not see. But these experiences also taught me to observe myself honestly, recognize problems early, and build systems and relationships that allowed me to keep moving forward. I refused to let my circumstances determine the scale of my ambitions. At Tsinghua University, I pursued computer science despite initially being told by my family that I might not be capable of succeeding in such a difficult field. When research opportunities were not given to me, I created my own. I taught myself unfamiliar areas of AI, formulated research questions independently, and eventually published three first-author papers, earned a GPA approaching 3.9/4.0, received major academic honors, conducted research at MIT, and received graduate offers from Harvard and Yale. I then learned to turn technical ability into something that could affect people beyond a classroom. At Pika, I led a 12-person multimodal AI R&D team working on generative video. Within three weeks, I shipped my first interactive demo, and within two months, our team delivered five major product releases. One reached more than 200,000 overseas views and 4,700 shares. The experience taught me that technology becomes meaningful when people can actually use it. My experiences also made me want to help people who face barriers similar to those I once faced. In 2022, I volunteered with Tsinghua University's Program Buddy Group, supporting underrepresented students interested in coding. I understood how powerful it can be when someone takes your potential seriously before you have the credentials to prove it. In the future, I want to use my talents in two ways: building AI technologies that solve meaningful real-world problems, and creating opportunities for people who lack strong support networks. I hope to mentor young students, expand access to technical education, and eventually build organizations where talented people are given the opportunity to grow regardless of their background. Being raised in a complicated family taught me that circumstances can shape a person without having to define them. I do not want my success to be an escape from my past. I want it to become a resource for others—to build technology, create opportunities, and become the kind of support I once needed myself.

The most important thing education gave me was not a degree. It was the ability to decide for myself what I was capable of. When I chose computer science, my family did not believe I could succeed in it. I was pushed toward another field because computer science was considered too difficult for me. I also watched my father teach my brother how to run his business while I was not given the same opportunity to learn from him. For a long time, other people seemed to have stronger expectations for what I could—or could not—become than I had for myself. Tsinghua University changed that. When I was not given the research opportunities I wanted, I learned not to wait for permission. I taught myself unfamiliar technical fields, formulated my own research questions, contacted researchers, and built projects independently. That persistence eventually led to a GPA approaching 3.9/4.0, three first-author papers, major academic honors including a Presidential Scholarship, research experience at MIT, and graduate offers from Harvard and Yale. Education also taught me that knowledge becomes most powerful when it can solve problems outside the classroom. During the 2020 Wuhan lockdown, an offline trading operation suddenly needed to move online. With no engineering team, external funding, or existing digital infrastructure, I built a digital operating system in 72 hours using Excel, DingTalk, and lightweight automation. It eventually processed more than 3,000 orders per day, reduced manual errors by 95%, shortened the order-to-cash cycle by 60%, and helped move 85% of revenue online within six months. The business ultimately exceeded ¥50 million in annual revenue and grew to 300 employees. These experiences shaped what I want my career to become. I want to work at the frontier of AI, particularly multimodal and world models that allow machines to understand physical and social environments. Ultimately, I hope to combine research, engineering, and entrepreneurship to build technologies that solve meaningful real-world problems. But I do not want the benefits of my education to stop with me. In 2022, I volunteered with Tsinghua University's Program Buddy Group, supporting underrepresented students interested in coding. I knew what it felt like to enter a technical field without everyone around you believing that you belonged there. I wanted to help other students find information, confidence, and opportunities earlier than I had. That is also how I think about this scholarship. The $1,000 would reduce the financial barrier to continuing my education and allow me to invest more time in research and technical development. Just as importantly, it would strengthen my ability to mentor other students and share the knowledge and opportunities I have gained. Education changed my future by giving me something no one else could permanently take away: the ability to build my own path. My goal is to use that ability to build a future not only for myself, but for the people who come after me.

My family once thought computer science was too difficult for me. Years later, I graduated from Tsinghua with a GPA approaching 3.9/4.0, published three first-author papers, received major academic honors, and earned graduate offers from Harvard and Yale. The distance between those two moments was built largely by creating opportunities that did not initially exist for me. When I chose computer science, my family pushed me toward another field because they did not believe I could succeed in it. I also watched my father teach my brother how to run his business while I was not given the same opportunity to learn from him. At university, I encountered a similar problem: when the research opportunities I wanted were not offered to me, I had to create my own path. I taught myself new technical areas, formulated research questions independently, reached out to researchers, and built projects from scratch. The results became my answer. I pursued multimodal AI research across vision, audio, language, and generative models, including research experiences at MIT. I published three first-author papers and received major academic recognition, including a Presidential Scholarship and Tsinghua's highest-level undergraduate honors. I was subsequently admitted to graduate programs at Harvard and Yale, although the cost of attending remained a significant financial barrier. My drive also extends beyond academia. During the 2020 Wuhan lockdown, I built a digital operating system for a trading operation in 72 hours, with no engineering team, external funding, or existing digital infrastructure. The system eventually processed more than 3,000 orders per day, reduced manual errors by 95%, shortened the order-to-cash cycle by 60%, and helped move 85% of revenue online within six months. The business ultimately exceeded ¥50 million in annual revenue and grew to 300 employees. Later, at Pika, I led a 12-person multimodal AI R&D team and helped ship five major product releases within two months. But one of the most meaningful ways I have used my opportunities is to help other students find theirs. In 2022, I volunteered with Tsinghua University's Program Buddy Group, working with underrepresented students interested in coding. I understood their uncertainty because I had experienced it myself. When you are the person others do not immediately expect to succeed, having someone who takes your ambition seriously can change the trajectory of your education. I wanted to be that person for someone else—to help students gain confidence, find resources, and see computer science as a field they could genuinely belong in. That experience also shapes how I would use this scholarship. The $1,000 would directly reduce the financial barrier to continuing my education and pursuing advanced AI research. But I would also treat it as an opportunity that should compound. It would allow me to spend more time on research, technical training, and mentorship rather than choosing opportunities based solely on cost. As I progress, I want to continue mentoring students from backgrounds that are underrepresented in computer science, sharing research resources, explaining how to enter technical fields, and helping students find opportunities that are often invisible to newcomers. My long-term goal is to work at the frontier of multimodal and world-model research and eventually build technologies and organizations with broad real-world impact. I want the opportunities I receive to create opportunities for others. I began my journey as a student whose potential was underestimated. I want my trajectory to end with other students being underestimated less often.

### 1. Please tell us a bit about yourself and what specifically motivates you. If you could do anything with your life, what do you imagine you’d do? The moment I understood what I wanted to do with my life was not in a laboratory. It was during the Wuhan lockdown, when a traditional trading business I was helping suddenly had to move online. There was no engineering team, no digital infrastructure, and no time to wait for someone else to solve the problem. So I built the system myself. Within 72 hours, I turned fragmented manual workflows into an online operating system that eventually processed more than 3,000 orders a day. That experience taught me something that has guided me ever since: when a system breaks, I want to be the person who figures out how to rebuild it. I am motivated by difficult problems that sit between technology and the real world. I enjoy understanding how something works, finding the bottleneck that prevents it from working better, and then building a solution that can scale beyond one person. If I could do anything with my life, I would build technologies that fundamentally expand what people and organizations can accomplish. I want to work at the frontier of AI, but I do not want to stop at research papers or prototypes. I want to build systems that people actually use—systems that transform industries, create opportunities, and eventually help intelligent machines understand and operate in the physical world. Ultimately, I want my career to follow one principle: turn impossible problems into working systems. ### 2. What excites you about STEM? How do you feel you could make a positive impact on the world through a job in information technology? What excites me most about STEM is that it turns curiosity into something tangible. An idea can begin as a question, become an algorithm, and eventually become a system capable of affecting thousands or even millions of people. I experienced this firsthand when I built digital infrastructure for a traditional business during the Wuhan lockdown. With no engineering team or external funding, I used simple technical tools to rebuild critical business workflows. The resulting system processed more than 3,000 orders per day, reduced manual errors by 95%, and helped move 85% of the company’s revenue online within six months. I learned that technology does not need to be complicated to be transformative; it needs to solve the right problem. Later, at Pika, I led a 12-person multimodal AI R&D team developing generative video products. I saw how quickly advances in machine learning could move from research ideas to products reaching a global audience. Today, I am exploring world models and embodied intelligence, with the goal of bringing AI beyond screens and into the physical world. Through information technology, I hope to make advanced capabilities more accessible and useful. I want to build tools that help small organizations operate more efficiently, enable people to accomplish tasks previously beyond their reach, and eventually allow intelligent systems to assist people in the physical environments where they live and work. For me, STEM is ultimately about leverage: using knowledge and technology to create impact far beyond the scale of one individual. ### 3. Please describe the greatest challenge you’ve had to overcome and how you managed to get past it. My greatest challenge was learning to take responsibility when there was no clear path forward. In 2020, during the Wuhan lockdown, the traditional trading business I was helping faced an existential problem. Offline operations had effectively stopped, customers needed to move online, and the company had no engineering team, digital infrastructure, or external funding. I was young, had never built a business operating system before, and had very little time to figure it out. My first reaction was to realize that waiting for the “right” resources was itself the problem. I broke the situation into the smallest solvable pieces: how orders entered the system, how inventory was tracked, how customers communicated, and how payments were settled. I learned what I did not know, built simple tools, tested them immediately, and kept replacing manual processes one by one. Within 72 hours, I had built the first functioning digital workflow. Over time, it grew into a system processing more than 3,000 orders per day, reducing manual errors by 95% and shortening the order-to-cash cycle by 60%. Within six months, 85% of revenue had moved online. The experience changed how I respond to uncertainty. I no longer see a lack of resources as a reason to stop. I see it as an engineering constraint. That mindset has stayed with me in research, entrepreneurship, and leadership: define the problem, build the first version, learn from reality, and keep moving.

When Wuhan went into lockdown in 2020, I learned that contributing to a community does not always mean volunteering in a visible way. Sometimes, it means making sure that an essential system continues to work when everyone else is struggling. My family operated a traditional food and supply-chain business in Wuhan. When the city shut down, our offline operations were suddenly disrupted, while supermarkets and local customers still depended on us for essential goods. There was no engineering team, no digital infrastructure, and no external funding. I realized that if we could not reorganize the business quickly, the consequences would extend beyond our company: employees could lose their livelihoods, customers could lose access to supplies, and local stores could face shortages. I decided to build the infrastructure myself. Within 72 hours, I rebuilt our core operations using Excel, lightweight automation, and custom scripts. I digitized order management, inventory tracking, customer communication, and settlement processes, allowing the business to continue operating despite the lockdown. The system eventually processed more than 3,000 orders per day, reduced manual errors by 95%, and enabled 85% of our revenue to move online within six months. More importantly to me, it helped keep our employees working and our customers supplied during an extraordinarily uncertain period. That experience changed how I think about technology and community. I realized that the most meaningful technology is not necessarily the most sophisticated technology. Its value comes from whether it makes people more resilient when circumstances become difficult. I also learned that I want to build systems that amplify the capacity of ordinary people rather than simply automate isolated tasks. Since then, I have pursued AI research and product development, working on multimodal generative systems and now exploring world models and embodied intelligence. I see these technologies as an opportunity to extend the lesson I learned in Wuhan: intelligent systems should ultimately operate in the real world and help people navigate uncertainty. In the future, I want to contribute by building AI systems that make critical real-world services more accessible, resilient, and responsive—particularly for communities and organizations that do not have the resources to build advanced technology themselves. I hope to combine engineering, entrepreneurship, and research to create tools that can move from laboratories into everyday life. The lockdown taught me that community contribution is not measured only by how many people you can reach directly. Sometimes it begins with one broken system, one urgent problem, and the decision to take responsibility for fixing it. I want to keep doing exactly that—building systems that allow people and communities to withstand disruption and create opportunities for others.