Interview Trainer AI Chatbot

A powerful desktop tool to practice technical interview questions with real-time AI grading and feedback.

Built for Spark, Data Engineering, Data Modeling, System Design, and other technical roles.
You answer → GPT-4o-mini scores you (0–10) + gives detailed feedback + improvement tips + bonus suggestions.

---

✨ Features

• Random question engine (no repeats in one session)
• Category filter – practice only the topics you want
• Smart "Pick a Question" window with live search bar (type any keyword → real-time filtering, case-insensitive)
• Clean question display format: 57 — Data Modeling: What is Data Vault modeling? (no more "Q57")
• Questions sorted numerically by ID
• Live timer per question + Reset Timer button
• "Read more" button – opens external link or lets you add one instantly
• Smart AI feedback with:
  • Colored score (green/orange/red)
  • Bold sections: What You Did Well, How to Improve, Bonus Points Suggestions
  • Scrollable content + separate Reference/Model Answer section
• Retry This Question (creates a new attempt record)
• History viewer:
  • All history
  • Only current question history
  • Sort by score, time, or ID
• Debug mode (prints full prompt + GPT response to console)
• Professional Question Builder (builder.py) – easy GUI to add/edit questions
• Loading spinner while waiting for AI
• History saved permanently in history.jsonl

---

Screenshots

1. Sample question display

2. Display AI-generated result after a submission

3. Show record history

4. Builder tool (for expanding the bank)

---

Folder Structure

interview_trainer/

├── assets/

│ ├── questions.json ← Your question bank

│ ├── history.jsonl ← All past attempts + scores

│ ├── grading_prompt.txt ← Editable AI prompt template

├── main.py ← Main interview trainer

├── builder.py ← Question editor tool

├── .env ← OpenAI API key

├── run.bat ← Double-click to start

├── requirements.txt

└── README.md

---

Installation

Enter this to the console to download all required lib: pip install -r requirements.txt

---

Setup

1. .env file (in root folder) -> put your API key on this file

• OPENAI_API_KEY=sk-proj-XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

2. grading_prompt.txt (must contain these placeholders)

Question from category: <<QUESTION_CATEGORY>> Question: <<QUESTION_TEXT>> User answer: <<ANSWER_TEXT>>

3. questions.json example JSON Question Block

[ { "qid": "1", "category": "Spark Architecture", "question": "What is the High-level Architecture of Spark?", "quick_answer": "Driver + Cluster Manager + Executors...", "short_answer": "Full detailed answer here...", "link": "https://spark.apache.org/docs/latest/" } ]

---

How to Use

1. Start the Trainer: Double-click run.bat (or run python main.py)
2. Add/Edit Questions: run builder.py and enter the question content.

• Browse to your questions.json
• Fill fields (Question ID auto-assigns if blank)
• Click Save / Append

3. Practice

• Click New Question
• Type your answer
• Click SUBMIT → Get AI Feedback
• Use Retry This Question or Next Question

4. View History

• Display History (All): To display the historical list of all submission, data retrived from history.JSONL file.
• Display This Question: Similiar to Display History, but this filters only that exact question.

---

Keyboard Shortcuts

• Ctrl + Z → Undo in answer box
• Ctrl + V → Paste

---

Troubleshooting

1. No questions appear

→ Check questions.json is a valid array and path is correct.

2. Feedback window looks broken

→ Resize the window slightly — it auto-adjusts.

3. API rate limit / cost

→ Using gpt-4o-mini (fast + very cheap).

---

Author

Built by Hoang Anh Tuan

Hanoi, Vietnam

Feel free to customize the prompt, add more categories, or extend the history viewer.

Happy practicing!

May your next interview be a 10/10 🔥