
“Writing Task 2 was the leak. The AI feedback flagged 'underdeveloped argument' on every essay until I rebuilt my structure with Yilin. Then it stopped flagging it.”

Every name, score, and university below is published with consent. Quotes in Mandarin show the original; English translation sits beneath. Small early-cohort numbers — every story is a real student.
I came to Melbourne in 2021. By third year I could write a 3,000-word essay over a weekend, but I'd plateaued at IELTS 6.5 overall. I'd done two prep courses. Both gave me handouts and ten-question quizzes. Neither put me through a full, timed exam.
The mock-test bank was the part I'd been missing. Full-length papers, all four sections, exam timing. The AI feedback after each one wasn't a percentage — it was a per-criterion band breakdown that read like an examiner's report. 'Task achievement: 7. You don't develop your second body paragraph.' That was specific. That was actionable.
Tutoring was once a week, in person at the Clayton studio. Yilin watched me write live, told me where I was losing marks for cohesion, and rebuilt my paragraph structure in two sessions. Speaking was a separate beast — we ran Part 2 cue cards on a kitchen timer until I stopped self-correcting mid-sentence.
I sat the test the fourth time. 8.0 overall, 7.5 minimum across all four. Lodged 189. The points were there.
“I'd taken IELTS three times. Same overall, every time. The mock-test bank was the unlock — I learned where the time was leaking. Tutoring fixed the writing in two sessions. Speaking was the last 0.5.”— Lin Y., Bachelor of Commerce

“Writing Task 2 was the leak. The AI feedback flagged 'underdeveloped argument' on every essay until I rebuilt my structure with Yilin. Then it stopped flagging it.”


“I sat General. Letter-writing was alien at first. Two in-person sessions unpacked every letter pattern. 7.0 first try.”
原文 · “考的是 General,写信的题型一开始我完全不会。两次面授课就把所有 letter 模板讲清楚了。雅思 G 7.0 一次过。”


“Six weeks. Twice-weekly tutoring, four mocks a week with AI breakdown. Describe-image was my weak spot — by week three the algorithm stopped docking me on content.”


“Computer-marked exams are won on tempo, not vocabulary. Tempo coaching is the part nobody else does — Yilin literally tapped a metronome on the desk.”

Pulled directly from our WhatsApp group for early students and our 小红书 channel, with permission. No screenshots — these are the students’ own words. Translations follow the original.
“Yilin rewrote my Task 2 live in front of me — three paragraphs gone. Hearing it in person beats a thousand YouTube videos.”
原文 · “今天 tutor 当场改我的 Task 2, 三段全部重写. 听到耳朵里, 比看视频管用一万倍.”
“PTE describe-image — 30 seconds, no filler. The AI timestamped every one of my pauses. Absurdly useful.”
原文 · “PTE describe-image 30 秒不会废话. AI 给我每一个停顿都标了时间戳. 离谱地有用.”
“The mock-test interface is the actual IELTS interface. Sitting the real one felt like the eighth practice round, not the first real one.”
One name per university so you can find the closest mirror to your own situation. Click through to read the full story.
1 student“Writing Task 2 was the leak. The AI feedback flagged 'underdeveloped argument' on every essay until I rebuilt my structure with Yilin. Then it stopped flagging it.”
1 student“I sat General. Letter-writing was alien at first. Two in-person sessions unpacked every letter pattern. 7.0 first try.”
1 student“Six weeks. Twice-weekly tutoring, four mocks a week with AI breakdown. Describe-image was my weak spot — by week three the algorithm stopped docking me on content.”
1 student“Computer-marked exams are won on tempo, not vocabulary. Tempo coaching is the part nobody else does — Yilin literally tapped a metronome on the desk.”
雅思 Writing T2 · 7 分作文逐句拆
PTE describe-image · 30 秒模板的真相
雅思 Listening Section 4 · 失分集中点
PTE 79+ vs 雅思 7+ · 该考哪个
Free 30-minute assessment. We diagnose your band, set a target, match a tutor. 30-day full refund if you see no lift.