VU Internship Program

AI Tools in
Mini‑Project Development

A hands-on guide to using AI for guided work, resolving technical questions, and structuring professional reports and slides.

Guided Work Technical Q&A Report Writing Slide Design Project Checklist
Unit 1 · Introduction

What AI tools can do
for your mini-project

AI assistants (ChatGPT, Claude, Gemini, Copilot) are not magic — but they are powerful collaborators when used with the right prompting strategy. This class shows you exactly how to integrate AI at each stage of your internship mini-project.

🧭

Guided Work

AI as a step-by-step mentor: breaking down complex tasks, suggesting roadmaps, and reviewing your progress.

🔧

Technical Q&A

Resolving bugs, understanding algorithms, and getting explanations tailored to your current knowledge level.

📄

Reports & Slides

Structuring professional documents, generating outlines, improving language, and building slide narratives.

⚡ Ground Rule
AI should amplify your thinking — not replace it. Always understand what the AI produces before submitting it. Use AI as a reviewer, planner, and explainer; keep the core ideas your own.
Unit 2 · Guided Work

Using AI as a project mentor

The most productive AI use is treating it like a patient senior colleague. You explain what you're building and ask it to help you think — not just execute. The key is the prompt quality.

Workflow stages where AI helps most

Define your project scope clearly

Before writing a single line of code, AI can help you understand what your project should and should not do. Paste your problem statement and ask for a scope analysis.

I am a CS student building a mini-project: [describe your project in 2-3 sentences]. My submission deadline is [N weeks] away and I can work [N hours] per week. Please help me define: (1) core features that are essential, (2) features I should skip for now, (3) realistic success criteria, and (4) the biggest risk that could derail this project.

🎯 Expected Output
AI will return a prioritized feature list (MoSCoW or similar), clear exclusions, and a risk list. Use this as your project charter — show it to your supervisor.

Build a realistic project plan

Once scope is clear, convert your scope into a week-by-week plan. AI is excellent at breaking down work into tasks with time estimates.

Based on this project scope: [paste scope from previous step]. I have [N] weeks. Create a Gantt-style breakdown showing: week number, main task, deliverable at end of week, and which technical skills I'll need. Format as a markdown table.

💡 Pro Tip
Ask AI to add a "buffer week" before your deadline. Projects almost always take longer than expected. Ask: "Add one contingency week and tell me what should go in it."

Get guided through unfamiliar tasks

When you hit a task you've never done before, don't guess — ask AI to walk you through it step by step, at your level.

I need to implement [specific feature/algorithm] in my project. I have intermediate knowledge of [Python/Java/etc.] but I've never done this before. Please: (1) explain the concept in simple terms, (2) give me a step-by-step implementation plan before any code, (3) only show code for step 1 first and wait for me to confirm I understand before continuing.

⚠️ Critical Habit
Always ask AI to "explain step 1 before showing code for step 2." This forces you to actually understand what's happening — which you'll need when your supervisor asks questions.

Get your work reviewed by AI

Before submitting code or a document draft to your supervisor, run it through AI for a structured review.

Please review the following [code/document section] as if you were a CS professor evaluating an MTech student's mini-project. Identify: (1) correctness issues, (2) efficiency improvements, (3) missing edge cases, (4) documentation gaps. Be specific and critical — I want to improve before submission.

[paste your work here]

Summarize and reflect on your project

At the end of your project, AI can help you articulate what you learned — which is crucial for viva voce (oral defence) preparation.

Here is a summary of my mini-project: [brief description]. Technologies used: [list]. Main challenge I solved: [describe]. Please generate 10 likely viva voce questions my examiner might ask, ranging from basic to advanced, and give me the key points I should cover in each answer.

🎓 Exam Prep
Practice answering these questions out loud before your viva. AI can also simulate the examiner: "Ask me these questions one by one and critique my answers."

Prompting techniques that actually work

🎭
Role Assignment
Begin prompts with "Act as an experienced software engineer reviewing a student project..." — giving AI a role dramatically improves response quality and focus.
high impact
🔢
Numbered Output Format
Ask for numbered lists, tables, or structured formats. "Give me exactly 5 steps, numbered, with a one-line summary and a code snippet for each." Vague prompts get vague answers.
structured
🪜
Progressive Disclosure
Ask AI to explain step 1 and pause. Confirm you understand, then ask for step 2. This prevents cognitive overload and ensures genuine learning rather than copy-paste.
learning-first
🔁
Iterative Refinement
First draft is never final. Ask: "What's wrong with this?" then "Now improve it" then "What are three alternative approaches?" Each loop improves quality significantly.
best practice
Unit 3 · Technical Questions

Resolving technical problems
with AI

AI is extraordinarily good at debugging, explaining errors, and teaching concepts. But the quality of the answer depends entirely on how you frame the question. Here are real scenarios you'll encounter.

Interactive Demo — click a scenario
My Python code throws: KeyError: 'user_id' when I call data['user_id'], but I printed data and it shows the key is there. Here is my code: [code block]. What is wrong and how do I fix it?
This is a very common issue with whitespace or encoding. The key "user_id" in your dict might actually be "user_id " (trailing space) or " user_id" (leading space), which is invisible when printed. Try: print(repr(list(data.keys()))) to see exact key strings. Also check if data comes from JSON parsing with extra whitespace, or if the dict was built from a CSV where headers had spaces. Fix: use data.get('user_id') which returns None instead of crashing, or strip all keys with {k.strip(): v for k, v in data.items()}.

The anatomy of a great technical question

TEMPLATE — copy and fill in your values

// CONTEXT
I am building: [project description]
Language/framework: [Python 3.11 / Java 17 / React 18 / etc.]

// PROBLEM
What I am trying to do: [describe intended behaviour]
What actually happens: [exact error message or wrong output]
When this happens: [always / sometimes / only with large inputs / etc.]

// CODE
[paste the relevant 10-30 lines only — not your entire file]

// WHAT I ALREADY TRIED
[list 1-3 things you already attempted]

Please explain WHY this error occurs first, then give the fix with an explanation of what changed.

🏆 Golden Rule
Never paste your entire 500-line file. AI works best on focused 10–30 line snippets. If the bug is complex, describe the surrounding context in words — don't dump code.

How to ask AI to teach you a concept

When you don't understand a concept (hashing, recursion, REST APIs, etc.) use this scaffolded learning prompt:

Explain [concept name] to me using this exact structure:
1. A one-sentence definition (no jargon)
2. A real-world analogy (not a computing analogy)
3. A tiny working code example in [Python/Java] (under 15 lines)
4. One common mistake students make and how to avoid it
5. One follow-up question I should ask next to go deeper

My current knowledge level: [beginner / intermediate / advanced]. Adjust your language accordingly.

Unit 4 · Report Writing

Structuring professional reports
with AI support

A well-structured report is as important as the project itself. AI can help you plan the structure, improve your language, and ensure completeness — while the actual content and insights must come from you.

Report Structure Builder

Select report type
Mark completed sections
AI prompt for selected report type
Select a section in the checklist and click "Generate AI Prompt" to see a tailored prompt for that section...

Language improvement prompts

Use these after drafting each section to improve academic quality:

✏️
Clarity Check
Paste a paragraph and ask: "Rewrite this for an academic technical report. Remove vague words. Make every sentence specific. Keep the same meaning." Compare before/after to learn the differences.
language
🔍
Completeness Check
Ask: "I have written the Methodology section. What important details am I likely missing for a CS mini-project report? List them as questions I should answer."
structure
📊
Results Narration
Give AI your numbers/charts description and ask: "Write a 150-word results narration that interprets these findings without overstating conclusions. Academic tone."
analysis
📝
Abstract Generator
After writing your full report: "Based on these key points from my report: [list 5-7 bullets], write a 150-word abstract following the IMRaD structure (Intro, Methods, Results, Discussion)."
summary
⚠️ Academic Integrity
AI-generated text must be treated as a first draft that you substantially revise. Do not submit AI output verbatim. Your institution's policy applies — use AI to support your writing, not replace it. Always cite methodology and tools used if required.
Unit 5 · Presentations

Building compelling slides
with AI guidance

Presentation slides are a different medium from reports — they must convey ideas visually and support speech. AI helps you structure the narrative, write slide titles, and craft speaker notes.

01Title slide
02Problem statement
03Approach / method
04Demo / results
05Conclusion
06Q&A / backup

The master slides prompt

I need to create a [N]-minute presentation for a [viva / conference / class demo]. My project: [2-sentence description]. Audience: [professors / peers / industry judges].

Please generate:
1. A slide-by-slide outline (slide number, title, 3-4 bullet points of content, one visual suggestion)
2. For each slide: one sentence the speaker should say when that slide appears (the "hook")
3. Three slides I should NOT include and why they would weaken the presentation
4. The single most important thing to emphasise to make a strong impression

Unit 6 · Project Checklist

Your AI-assisted
mini-project checklist

Use this interactive checklist throughout your project. Tick items as you complete them. Each item maps to an AI workflow covered in this tutorial.

0 / 20 completed
🚀 You're ready
Once this checklist is 100% complete, your mini-project is submission-ready. Each of the AI workflows in this tutorial maps to a checklist item — revisit the relevant unit whenever you're stuck.