Stage 1: Hacking the Vocab

As an international student at The Fessenden School, the SSAT vocabulary section felt like a wall. Lists of obscure words, no context, and the pressure of timed sections—rote memorization was not only boring; it didn’t stick. I needed a way to learn words in real sentences and track what I actually got wrong so I could improve systematically.

The Geeky Solution

Rote memorization doesn’t work for me. I learn better when words appear inside the stuff I’m already reading—novels, articles, textbooks. So I built a pipeline that turns my own reading into a personalized vocab lab.

Here’s the idea: a Python script takes my reading materials (for example, the full text of Doctor Zhivago) and pulls out challenging vocabulary in context. Then it sends those passages to the Gemini API, which generates contextual quizzes: fill-in-the-blank, synonym choices, or short-answer questions tied to how the word is used in the book. I don’t just see a word and a definition; I see the sentence and have to reason about meaning.

Every quiz response is logged—right answers, wrong answers, and which words I keep missing. That becomes a dataset: error rates per word and over time. I can spot my weak words, retest them, and watch the error rate go down. It’s vocab prep as a small data project: same stack I use for other ideas (Python, an LLM API, a bit of persistence), but applied directly to the SSAT and TOEFL. I’m documenting this approach so other public school students can adapt it to their own reading list and goals.