Ruitong Tang

The problem

The company had long articles, research reports and YouTube videos, but almost nothing made for TikTok, where many prospective students actually are.

What I did

I built a repeatable workflow: pick the source, choose video or carousel by its structure, write the script and cut sheet, produce it in CapCut and Canva, and give every post a call to action that lands on the matching page.

The result

66 videos and carousels in six months, each with a working link to a specific page on the site.

The problem

Each scholarship cycle has many steps and many people, and one wrong email or an early social post lands on a real student's name.

What I did

I ran the whole chain: selection, notifications, interviews with my manager, a scored recommendation for the supervisor, then publishing in a fixed order: written acceptance first, then the other finalists, then social, then the article.

The result

8 campaigns (6 of them end to end) with partners such as IAESTE and Girls on Campus, 20+ finalists interviewed from 15+ countries, 12 articles published, and no announcement ever out of order.

The problem

Reading every application by hand took about two weeks per scholarship, and it was easy to miss things in long, similar-looking essays.

What I did

I wrote a two-round Python tool with AI-assisted development. Round one checks eligibility and writes a reason on every row. Round two scores how specific each essay is. Unclear cases go to a person, and the supervisor makes every final call. No AI decides anything about an applicant.

The result

Reused across 5 scholarships. Two weeks of reading became a run measured in seconds, with human review after. My manager said publicly that it saved the team hours every week.

The problem

Interviews only served to pick a winner, and the brand still did not know why students choose a country or a university, or what nearly stops them.

What I did

I rebuilt the interview around three questions (why this country, why this university, what nearly stopped you), and wrote a script that summarises application data into counts only, inside the company's own Microsoft 365 system, so no individual data ever leaves it.

The result

One interview now feeds three things: the winner story, social content and a cycle-end audience brief. Individual data stays put; only totals come out.

The problem

Brands watching a controversy usually count mentions. But on a platform ranked by likes, a handful of comments carry most of what anyone actually reads, so counting everything equally describes a conversation nobody had. I wanted to know how a shared opinion forms in that gap, using a case that crossed three markets at once.

What I did

I collected 5,289 comments under 219 Weibo posts and sampled by visibility instead of volume: every comment above 500 likes (246 of them, carrying 86.9% of all the likes), plus a random tail of 200 so I could still say something about the quiet majority. I coded 446 comments, close-read 29, and wrote down all 28 borderline judgement calls so another person could check the work. I did not clean the corpus first, because filtering to the relevant comments would have turned rarity into something I had created myself.

The result

Passed public defence at Stockholm University, August 2026. The two vocabularies moved in opposite directions: nation words appeared in 260 comments, class words in only 98, yet class-framed comments drew about three times the likes each (699 against 231). Counting would have called the argument nationalist. Weighting by what people saw showed it was settled by class, and settled by renaming the thing rather than by arguing about it.