How I Replaced ChatGPT's Empty "Great!" with a Real Interview Trainer Using Claude Code

ChatGPT Always Says "Great!" But I Don't Remember Anything—Building a Real Interview Teacher with Claude Code

A Pain Point for Everyone

Have you ever encountered this when using ChatGPT / Claude to prepare for an interview:

• You give a clearly flawed answer, it responds "Great answer!" and moves on to the next question.
• Every time you start a new conversation, it doesn't remember what you learned yesterday, and you have to introduce yourself from scratch.
• You've memorized a bunch of stock answers, but don't know when to review them, and forget everything after a couple of days.
• It explains a concept to you, and you think you understand it after reading, but when it's time for the interview, you can't articulate it.

The root cause is simple: ChatGPT is a polite conversational partner, not a strict teacher.

Its operating principle is "keep the user satisfied," not "help the user learn." It thinks whatever you say is good because it's afraid you'll be unhappy.

Is There a Way to Make AI a Strict Teacher?

Yes. I did it using Claude Code (Anthropic's CLI programming tool).

Claude Code has a unique capability: Rule files (CLAUDE.md) + a persistent memory system. Simply put, you can "program" the AI's behavior by writing rule files, and these rules are automatically loaded in every conversation.

I applied this capability to interview preparation.

What I Did

I wrote a set of configuration files with two core components:

1. Strict Learning Process (rules/interview.md)

Rules:
1. Mix questions (from 3-4 modules, not just one topic)
2. Correct answer → Brief confirmation + related knowledge points → Next question
3. Wrong answer → Must explain (including reference URL) → Give a verification question → Only then move to a new question
4. No skipping steps, cannot go directly to the next question after answering.

Difference from standard AI: The rules are fixed, not decided by the AI itself. It can't lazily say "Great" because the rule file tells it it must complete the explanation → verification process.

2. Spaced Repetition Review Schedule (knowledge/interview_tracker.md)

Knowledge Point	First Learned	D+1	D+2	D+4	D+7	D+15	Status
Smart Pointers	5/29	✅	✅				Mastered
Virtual Function Table	5/29	❌					Weak

At the start of each session, Claude automatically checks which knowledge points are due for review and prioritizes review questions. You don't need to remember when to review yourself.

3. Daily Wrap-up Verification

Say "That's it for today" to trigger the wrap-up process:

1. Claude provides a daily learning outline.
2. You recall based on the outline and give your own understanding.
3. Claude judges whether you truly understand or are just memorizing answers.
4. Updates the review schedule.

This is the AI version of the Feynman Technique—it's not about you thinking you understand, but about you being able to explain it, with the AI as the judge.

Comparison of Effects

Capability	ChatGPT Interview	Anki	NotebookLM	SaaS Platforms	This Project
Cross-session memory	None	Deck-level only	None	Partial	Yes
Spaced repetition	None	Yes	None	None	Yes
Wrong answer→Explain→Verify	None	None (just gives answer)	None	Partial	Yes
Native question generation from project code	Need to manually paste code	N/A	Need to upload files	Not supported	Native support
Additional costs	None	None	None	\$29-300/time	No extra cost

Core difference: Other tools either have memory without understanding (Anki), or understanding without memory (ChatGPT), or neither (SaaS). Claude Code's rule files + memory system is the only solution that combines both.

The Ultimate Killer Feature: Starting from Your Project Code

All interview tools on the market follow a "I ask, you answer" model.

This project is different: It asks follow-up questions based on your own code.

For example, my project has a BoundedQueue (a bounded queue) that uses two std::condition_variables. When preparing for an interview, Claude doesn't ask you "Please explain condition_variable," but instead asks:

"Your BoundedQueue uses two condition_variables (not_full_ and not_empty_). Could it work with just one?"

This kind of question forces you to think about the principles from code you've actually written. The knowledge is anchored, not just memorized stock answers.

Quick Start

1. Fork the Repository

gh repo fork happiness-cheng/ai-interview-engine --clone

2. Open with Claude Code

cd ai-interview-engine
claude

3. Say Your First Sentence

I want to prepare for a C++ backend development interview, targeting an internship at a top company. Start reviewing.

That's it. Claude will read the rule files and start the first round of questions.

4. Switch to Your Own Tech Stack

The repository comes with a sample question bank for C++ backend. If you use Java / Go / Front-end, just replace it with the blank template:

knowledge/TEMPLATE.md → Copy to interview_tracker.md → Fill in your knowledge points

My Experience

I am a sophomore student preparing for the 2026 fall recruitment internship. Using this system to prepare for C++ backend interviews, the most noticeable feelings are:

1. Less anxiety — No need to plan "what to review today" yourself; the system automatically schedules based on the forgetting curve.
2. Actually remembered — The "Wrong answer → Explain → Verify" process forces you to truly understand, not just skim over it.
3. Confidence in my project — Follow-up questions based on your own code let you clearly explain the "why" behind every design decision in your project.

This Is Not Just an Interview Tool

The underlying engine is universal: Question → Answer → Judge → Explain → Verify → Forgetting Curve Scheduling.

Interviews are just one scenario. You can use it to prepare for:

• Technical interviews (currently supported)
• Career change/job hopping review
• Final exams (same principle, just swap the knowledge base)

Open Source Address

GitHub: https://github.com/happiness-cheng/ai-interview-engine

If it helps you, give it a star. If you have suggestions for improvement, feel free to open an issue.