Tracking what I needed to practise
While preparing for the GRE in 2023, I logged my answer times, the topics I found difficult and how challenging each question felt. The log gave me a way to choose what to practise next. Keeping it current took time.

Building a learning model for each learner
I built 31Memorize around spaced repetition, where reviews are spread over time. I developed a custom machine-learning model for each learner so the tool could use their own practice history to adapt what they reviewed.

Creating more questions from the same concept
I wrote templates that could generate new quantitative questions by changing their values while preserving the concept being tested. I also built rules for plausible wrong answers, such as results a learner might get by applying the wrong formula.

Presenting the work
I presented 31Memorize at the Alibaba Cloud Singapore AI Hackathon in March 2024. In August, the project reached the final evaluation stage for SMU's BIG incubator.
Evaluator feedback

“I really appreciate the passion of the candidate, who has continuously been learning from his past experiences success and failure.”
SMU - Business InnovationBIG incubator final-stage evaluation / 29 August 2024“Focused on the user pain points. Good initiative in learning language.”
Alibaba CloudEvaluation feedback / 8 March 2024