Hr ops · Production

Merpay iOS engineer uses AI-driven learning cycle with Cursor to master programming languages

The problem

As AI tools made it easy to generate working code without deep understanding ('vibe coding'), the author grew concerned about whether they truly grasped the code AI produced, and wanted a structured way to use AI to learn rather than just to produce output.

Workflow diagram · grounded in source
1
Step 1: AI creates learning plan
ai_action
“公式ドキュメントを参照し、体系的な学習計画を作成してもらいます。何を・いつ・どの順番で学ぶかが明確になります。”
2
Step 2: AI generates daily materials
ai_action
“学習プランに沿って、解説・サンプルコード・演習課題を含む教材を生成してもらいます。そのまま実行できるファイル形式で出力させるのがポイントです。”
3
Step 3: Human implements, AI reviews
feedback_loop
“教材の演習課題を自分で実装し、AIにコードレビューしてもらいます。ここが一番学びが深まるステップです。”
4
Step 4: AI tracks progress
ai_action
“学習プランからTODOリストを生成し、進捗を可視化します。次にやることが常に明確になり、継続しやすくなります。”
Reported outcome

The AI-driven learning cycle made it easier to continue studying, reduced the overhead of planning what to do next, and gave the author confidence that learning was sticking through output-centered practice.

Reported metrics
Learning continuity継続しやすくなりました
Learning retention確実に身につく実感
Ability to tackle new language domains新しい言語にも挑戦しやすくなり
Reported stack
Cursor
Source
https://engineering.mercari.com/blog/entry/20251223-ai-driven-learning/
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

The AI-driven learning cycle made it easier to continue studying, reduced the overhead of planning what to do next, and gave the author confidence that learning was sticking through output-centered practice.

What tools did this team use?

Cursor.

What results were reported?

Learning continuity: 継続しやすくなりました; Learning retention: 確実に身につく実感; Ability to tackle new language domains: 新しい言語にも挑戦しやすくなり (source-reported, not independently verified).

How is this hr ops AI workflow structured?

Step 1: AI creates learning plan → Step 2: AI generates daily materials → Step 3: Human implements, AI reviews → Step 4: AI tracks progress.