I chose to take part in an exchange program at Xi’an Jiaotong University during a period when I could have gone to Beijing, Shanghai, Shenzhen, Europe, or several top-tier institutions with existing partnerships. After spending years moving between Tokyo, Taipei, Seoul, and San Francisco, I wanted to understand how technology, economics, and culture evolve outside the global first-tier circuit. Xi’an offered that vantage point. It is a historic city anchored by one of China’s most respected research universities, and it represented a very different kind of frontier for me to observe.
Around the same time, I received another offer: an opportunity to join a top-tier research lab in Hong Kong. That path would have placed me firmly toward academia, with a clear trajectory and strong institutional support. Choosing Xi’an, however, meant choosing a broader lens—stepping into a different cultural and economic context and forming my own understanding rather than following a predefined academic route. In the end, that was the direction I committed to.
Xi'An Jiaotong University
Xi’an Jiaotong University is one of China’s leading research universities and a member of the national C9 League. It is renowned for its strengths in engineering, energy systems, electrical and power sciences, materials, and other disciplines that anchor the country’s advanced industrial and aerospace sectors.
Multiple fields are included in national and provincial “Double First-Class” programs, and the university has long undertaken major state-supported research while maintaining close collaboration with energy, equipment manufacturing, and high-end industrial partners. Its notable alumni include Qian Xuesen, a foundational figure in modern aerospace.
The School of Management
Although I had the freedom to enroll in any school across the university, I chose the School of Management for a very specific reason: it is not a traditional business school. Two departments in particular stood out to me—Management Science and Industrial Engineering & Operations Management.
Both are deeply rooted in engineering and mathematical modeling, focusing on optimization, data-driven decision systems, complex networks, production and operations engineering, and system-level management.
These programs offered a rigor that aligned with my interest in analytical and engineering-oriented approaches to understanding organizations and large-scale systems. Most of the courses I took came from these two departments, where management is treated not as administration, but as the study of systems, constraints, efficiency, and complex human-machine interactions.
The Curriculum
Advanced Management
I took a corporate governance module that used one of China’s most dramatic control-rights battles—the Vanke vs. Baoneng takeover case—to unpack how ownership structure, board design, incentives, and governance mechanisms shape the fate of a company. The class approached governance not as a legal formality, but as an economic design problem: who holds power, how that power is constrained, and how long-term value is protected when interests collide.
We traced how Vanke’s dispersed ownership and lack of a controlling shareholder opened the door for Baoneng’s aggressive stake-building, triggering a governance crisis. From there, we examined competing theories—shareholder-primacy vs. stakeholder governance—and how they explain the tensions among founders, management, institutional investors, and retail shareholders. The course also highlighted tools like delegated voting mechanisms, equity-based incentives, and Alibaba’s partnership structure to illustrate how firms engineer control rights differently to preserve strategic continuity while satisfying modern governance expectations.
For me, the most meaningful takeaway was seeing governance as an architecture of incentives and constraints: a system that must balance power, embed checks-and-balances, and still enable innovation and long-term decision-making. It sharpened how I think about founder control, investor alignment, and how different governance designs can either stabilize a company—or make it vulnerable in moments of external pressure.
Project Investment Decision-Making and Management
I took a course on Project Investment Decision-Making and Management that focused on how to evaluate and finance large-scale industrial and infrastructure projects, such as port developments and high-speed rail lines. We treated each project as an investment decision: build cash-flow models, assess feasibility, design the financing structure, and then decide whether and how the project should move forward.
In this course I worked directly with the finance toolkit behind these decisions—building DCF models, running NPV, IRR, and ROI analysis, and using CAPM and the cost of capital to test whether a project is financially viable for private investors. For both private industrial projects and public infrastructure, we compared different capital structures and cash-flow profiles to see how they change the investment case.
My biggest takeaway, though, was from evaluating public projects. For private projects, the decision is mostly about financial and economic returns to investors. For public projects, you have to quantify much broader impacts—using economic indicators and models to turn “social benefit” into numbers. For example, with a new rail line or port, we had to model how improved accessibility for remote areas, time savings, and regional development translate into measurable economic value. That process of turning messy, real-world public benefits into structured mathematical models fits how I like to think: using quantitative frameworks to make complex, high-stakes decisions more transparent and grounded.
Scalable Data Analytics and Machine Learning
This course was my real entry point into machine learning. Our professor had just returned from California—bringing back a curriculum that pushed us beyond Python basics into scalable data processing with tools like Scala, distributed data pipelines, and hands-on ML modeling.
For my project, I chose a well-known open movie-rating dataset spanning several decades. I used ML techniques to analyze film reviews, genre patterns, and audience scores, but the most surprising insights came from my own interpretation—not the model’s. After clustering and modeling trends, I noticed that films aligned with contemporaneous social movements (such as feminist waves or political moments) consistently received higher ratings. The model surfaced patterns, but the deeper meaning—the social alignment behind those patterns—came from human reasoning.
When I presented this, my professor highlighted exactly that distinction: ML can reveal structure, but humans give structure meaning. It was the first time I felt how complementary machine intelligence and human insight could be. The machine helped expose what might go viral or earn high scores, but my role was to understand why. That experience shaped how I think about data-driven creativity and how humans and models can collaborate in discovering non-obvious connections.