The Experience of Self-learning Driven by Interest

Unlike my peers majoring in computer science, I pursued a self-learning path based on my interests. While encountering some difficulties, the journey was also highly rewarding.

My first exposure to computer science was during my freshman year, where I learned C language - my first programming language. I found it fascinating and read the entire textbook thoroughly. As a result, I performed well in the course and scored 95 points. In my sophomore and junior years, I participated in mathematical modeling competitions and was introduced to neural networks. Although I didn’t fully understand them at the time, it sparked my curiosity to learn more. In my senior year, I heard about concepts like deep learning and artificial intelligence, and realized that they were built on neural networks, which I had not yet fully grasped. Driven by my curiosity, I planned to systematically learn the principles of these topics.

Throughout my graduate studies, starting from my senior year, I continued to explore computer science knowledge driven by my passion. I delved into Python, Machine Learning, Deep Learning, Java, C++, Data Structures and Algorithms. Along the way, I was introduced to lower-level concepts, so I started learning Operating Systems, Computer Networks, Database Principles, and Design Patterns. I also pursued many mathematics courses, including Matrix Theory, Partial Differential Equation, and Computational Methods, with the aim of becoming a well-rounded engineering student with a strong foundation in mathematics and programming skills.

During my master’s degree, I chose to research on the intersection of deep learning and civil engineering. In a situation where the research group had no accumulation in this area, I had to figure things out on my own. I aspired to apply for a Ph.D. in computer science during my graduate studies, but struggled to find a clear path. I eventually decided to work instead. However, as I gained more exposure to information, I discovered that a path to a Ph.D. could be pursued through internships at research labs to gain research experience and skills.

After graduation, I worked as a web backend developer at Baidu. As the entire backend system was a distributed system, I began learning about distributed systems and studied various open-source project architectures, such as Redis and Zookeeper. I also read papers on machine learning, including the Bert and GPT series.

Looking back on my entire experience of self-learning, I’ve realized that I’ve gained not only knowledge, but also the ability to select appropriate learning materials, effectively manage study time and energy, and develop the courage to overcome difficulties. With these skills, I am excited to continue exploring new areas of knowledge and challenging myself further.

My interest in both distributed systems and machine learning is the reason I want to work in MLSys. I believe that the infrastructure for large models is critical and promising. Additionally, this field is closer to industry and practical applications compared to other fields. The results of this field are more easily deployed in the development process and products, rather than just remaining on paper. This aligns with my research goals, which aim to have a positive impact on subsequent research in this field and its applications in industry.