01 · Preparation

I am completing a B.Eng. in Software Engineering at Donglan Institute of Technology, a fictional institution used only for this sample. Coursework in data structures, probability, linear algebra, databases and machine learning gave me the computational foundation to work with language data, while an elective in modern Chinese grammar taught me to examine what token-level features fail to capture.

02 · Evidence

In my capstone project, I assembled and documented a 12,000-item Chinese review corpus, compared a conditional random field baseline with a transformer-based classifier, and led error analysis across negation, aspect terms and mixed sentiment. The most valuable outcome was not a single score; it was learning to connect annotation choices, model assumptions and linguistic evidence in a reproducible workflow.

A subsequent internship with Qiming Data Studio, also fictional, placed that workflow in a product setting. I translated recurring customer-service questions into an intent taxonomy, reviewed mislabeled samples with operations colleagues and wrote concise model cards for two internal experiments. This experience showed me that useful language technology depends on communication across engineering, linguistics and the people represented by the data.

03 · Program fit

Stuttgart's English-taught program is a strong next step because it joins computer science and linguistics while emphasizing teamwork, laboratory practice and research competence. I am especially interested in using the program's research-oriented environment to deepen my understanding of linguistic analysis, statistical modelling and evaluation rather than treating NLP as a sequence of interchangeable tools.

My software-engineering background has prepared me for programming-intensive work, and I have planned focused study in formal linguistics before enrolment so that I can contribute from both sides of this interdisciplinary field. I will submit the current C1-level English evidence and academic records required for my own application cycle.

04 · Direction

After the master's degree, I intend to work on multilingual language systems for public-service and enterprise knowledge access in China. My long-term goal is to build evaluation practices that expose where a system fails across terminology, dialect and user context, so deployment decisions are based on traceable evidence rather than headline accuracy alone.

I would bring to the cohort experience in building annotated datasets, documenting experiments and explaining technical trade-offs to non-specialist collaborators. Just as importantly, I hope to test my assumptions with classmates whose languages and disciplinary backgrounds reveal different failure modes from the Chinese datasets I know best.