20 practical, difficulty-rated ways Computer Science students at Hobart and William Smith Colleges can use AI — each naming the exact tutorial it links to and what you’ll take away. Click any card to watch it.
Before you use these on graded work: check your professor’s policy on AI.
It differs by course, and these examples are study aids — not permission.
Easy
Explain an algorithm
Walk through how quicksort partitions and recurses. The linked tutorial covers a 10-step ChatGPT prompt system — standard definition, an explain-it-to-a-10-year-old version, real-life analogies, a simple diagram, then a comparison table — that you aim at any concept you're stuck on.
What you'll take awayA repeatable prompt sequence that turns a confusing idea into a definition, an analogy and a diagram you can check against your course notes.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Explain an algorithm — Walk through how quicksort partitions and recurses.
Start with a one-sentence definition, then explain it as if I'm 10, then give a real-world analogy. Finish by flagging anything scholars genuinely disagree about — and tell me what you're unsure of.
Add clear docstrings and inline comments to a function. The linked tutorial covers pasting a broken snippet into ChatGPT with the prompt 'scan the following code for potential problems' — it finds the bug (a divide-by-zero), explains why it happens and returns a fixed version.
What you'll take awayA debugging prompt formula, plus the habit of re-running the fixed code to confirm the error is actually gone.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Comment my code — Add clear docstrings and inline comments to a function.
Tell me what's wrong, explain why it happens, then give the corrected version. Walk through the fixed code line by line with a concrete input.
[PASTE YOUR CODE AND THE ERROR HERE]
Get the definition plus runtime versus compile-time cases in your language. The linked tutorial covers ChatGPT's Study Mode (type /study), an interactive tutor that guides you with questions step by step instead of just handing over the answer.
What you'll take awayHow to switch ChatGPT into tutor mode so it drills you on terms and ideas rather than doing the thinking for you.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Define polymorphism with a concrete example — Get the definition plus runtime versus compile-time cases in your language.
Define each term in one clear sentence, then give a concrete example of each in context. Once I've read them, quiz me on them one at a time and tell me what I got wrong and why.
Condense a library's API docs into key usage notes. The linked tutorial covers MacMost's copy-paste method — pull an article into reader view, paste it into ChatGPT after 'summarize this:', then ask for a set word count or a 10-bullet version; also how to grab a YouTube transcript to summarise a video.
What you'll take awayHow to compress a long reading or video into a summary of the exact length you need, and why pasting the source keeps it accurate.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Summarize documentation — Condense a library's API docs into key usage notes.
Give me (1) a three-sentence summary, (2) five bullet-point key claims, and (3) anything the author asserts without supporting evidence. Text below:
[PASTE THE TEXT HERE]
Ideas scoped to a single term, with the hardest technical risk named upfront. The linked tutorial covers three brainstorming methods in ChatGPT — a reusable custom-instructions 'brainstorm paradigm', an AI-diagram tree that fans one topic into dozens of branches, and a SWOT prompt to pressure-test an idea.
What you'll take awayThree prompt patterns you can reuse to go from a blank page to a categorised list of options, plus a way to stress-test the one you pick.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Brainstorm app ideas for a semester capstone — Ideas scoped to a single term, with the hardest technical risk named upfront.
Give me 10 options grouped into 3 themes, then mark the 3 most original and explain what makes them stronger.
Write setup and usage instructions for a small repo. The linked tutorial covers AI Academy's approach to long-form writing: never ask for the whole essay at once — break it into smaller prompts across two tabs and a document so you keep control of the argument.
What you'll take awayA workflow that automates the writing but not the thinking, so the argument stays yours and the draft stays checkable.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Draft a README — Write setup and usage instructions for a small repo.
Help me write this, don't write it for me. First ask me 3 questions about my argument, then propose a thesis and a section plan. Do not draft full paragraphs until I approve the plan.
Interpret a stack trace and suggest likely causes. The linked tutorial covers a 10-step ChatGPT prompt system — standard definition, an explain-it-to-a-10-year-old version, real-life analogies, a simple diagram, then a comparison table — that you aim at any concept you're stuck on.
What you'll take awayA repeatable prompt sequence that turns a confusing idea into a definition, an analogy and a diagram you can check against your course notes.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Explain an error message — Interpret a stack trace and suggest likely causes.
Start with a one-sentence definition, then explain it as if I'm 10, then give a real-world analogy. Finish by flagging anything scholars genuinely disagree about — and tell me what you're unsure of.
Translate pseudocode into a Python function draft. The linked tutorial covers pasting a broken snippet into ChatGPT with the prompt 'scan the following code for potential problems' — it finds the bug (a divide-by-zero), explains why it happens and returns a fixed version.
What you'll take awayA debugging prompt formula, plus the habit of re-running the fixed code to confirm the error is actually gone.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Convert pseudocode — Translate pseudocode into a Python function draft.
Tell me what's wrong, explain why it happens, then give the corrected version. Walk through the fixed code line by line with a concrete input.
[PASTE YOUR CODE AND THE ERROR HERE]
Turn Big-O complexities into quick review cards. The linked tutorial covers prompting ChatGPT for a set of question-and-answer flashcards on a topic, then exporting the set straight into Quizlet.
What you'll take awayA pipeline from topic to a study-ready Quizlet deck in a couple of minutes.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Create flashcards for Big-O complexities — Turn Big-O complexities into quick review cards.
Create 20 flashcards formatted as 'Question | Answer', one per line, so I can paste them straight into Quizlet or Anki. Keep each answer under 25 words.
Write an internship application letter for a software role. The linked tutorial covers drafting an email in ChatGPT and improving it through conversation — the first draft is deliberately generic, then detail is added turn by turn — plus what not to paste in for privacy.
What you'll take awayWhy the first draft always reads like AI, how to iterate it into your own voice, and to keep real names out of the prompt.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Draft a cover letter — Write an internship application letter for a software role.
Keep it under 150 words: polite, specific, with one clear ask. Give me two versions — one formal, one warmer. Don't include real names, ID numbers, or anything private.
Improve readability and structure while preserving behavior. The linked tutorial covers prompts for refactoring legacy code with ChatGPT — restructuring for readability and modernising outdated syntax without changing what the code does.
What you'll take awayRefactoring prompts that clean up messy code while preserving its behaviour.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Refactor code — Improve readability and structure while preserving behavior.
Tell me what's wrong, explain why it happens, then give the corrected version. Walk through the fixed code line by line with a concrete input.
[PASTE YOUR CODE AND THE ERROR HERE]
Weigh tradeoffs between hash maps and balanced trees. The linked tutorial covers a compare-and-contrast prompt in ChatGPT (demonstrated on Buddhism vs Hinduism) that asks for similarities and differences, then expands any point you want more depth on.
What you'll take awayA prompt for structured comparisons, and how to push ChatGPT to elaborate on the specific dimension you care about.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Compare data structures — Weigh tradeoffs between hash maps and balanced trees.
Give me a table of the key similarities and differences, then one short paragraph on which difference matters most and why.
Generate edge-case tests for a given function and review coverage. The linked tutorial covers giving ChatGPT a scenario and having it generate a full set of test cases from it.
What you'll take awayHow to turn a description of expected behaviour into structured test cases.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Write unit tests — Generate edge-case tests for a given function and review coverage.
Tell me what's wrong, explain why it happens, then give the corrected version. Walk through the fixed code line by line with a concrete input.
[PASTE YOUR CODE AND THE ERROR HERE]
Sketch components and data flow for a URL shortener. The linked tutorial covers pasting a broken snippet into ChatGPT with the prompt 'scan the following code for potential problems' — it finds the bug (a divide-by-zero), explains why it happens and returns a fixed version.
What you'll take awayA debugging prompt formula, plus the habit of re-running the fixed code to confirm the error is actually gone.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Plan a system design — Sketch components and data flow for a URL shortener.
Tell me what's wrong, explain why it happens, then give the corrected version. Walk through the fixed code line by line with a concrete input.
[PASTE YOUR CODE AND THE ERROR HERE]
Synthesize operating systems topics into an organized exam review. The linked tutorial covers building a custom study plan with ChatGPT using the Ultralearning why/what/how framework — benchmarking how others learn it, then a week-by-week plan with deliverables and grading rubrics.
What you'll take awayA prompt sequence that produces a personalised study schedule with checkable milestones, plus a pass where ChatGPT critiques the plan as an expert.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Build an operating-systems exam study guide — Synthesize operating systems topics into an organized exam review.
I have [X hours per week] and my exam is on [DATE]. Build a week-by-week plan with specific deliverables, and for each week give me a way to check I've actually learned it.
Use pandas to summarize and visualize a CSV, checking logic. The linked tutorial covers dragging a dataset straight into ChatGPT and prompting it to describe the data, produce descriptive statistics per column, and chart the columns you care about.
What you'll take awayHow to interrogate a spreadsheet conversationally, and why the 'analyzing' step means it ran real code rather than guessing.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Analyze a CSV dataset with pandas — Use pandas to summarize and visualize a CSV, checking logic.
I'm uploading my dataset. First describe what's in it and flag any data-quality problems, then answer my question, and show the code you ran so I can check it myself.
Diagnose a concurrency bug and verify the proposed fix. The linked tutorial covers pasting a broken snippet into ChatGPT with the prompt 'scan the following code for potential problems' — it finds the bug (a divide-by-zero), explains why it happens and returns a fixed version.
What you'll take awayA debugging prompt formula, plus the habit of re-running the fixed code to confirm the error is actually gone.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Debug complex code — Diagnose a concurrency bug and verify the proposed fix.
Tell me what's wrong, explain why it happens, then give the corrected version. Walk through the fixed code line by line with a concrete input.
[PASTE YOUR CODE AND THE ERROR HERE]
Translate pseudocode from research into working, tested code. The linked tutorial covers pasting a broken snippet into ChatGPT with the prompt 'scan the following code for potential problems' — it finds the bug (a divide-by-zero), explains why it happens and returns a fixed version.
What you'll take awayA debugging prompt formula, plus the habit of re-running the fixed code to confirm the error is actually gone.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Implement an algorithm from a paper — Translate pseudocode from research into working, tested code.
Tell me what's wrong, explain why it happens, then give the corrected version. Walk through the fixed code line by line with a concrete input.
[PASTE YOUR CODE AND THE ERROR HERE]
Design retrieval, embeddings, and prompts, then evaluate accuracy. The linked tutorial covers how retrieval-augmented generation works, then building a working RAG chatbot with Python, LangChain, Streamlit and a Pinecone vector database.
What you'll take awayThe architecture behind a document-grounded chatbot, and a working build you can adapt.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Build a custom RAG chatbot — Design retrieval, embeddings, and prompts, then evaluate accuracy.
Tell me what's wrong, explain why it happens, then give the corrected version. Walk through the fixed code line by line with a concrete input.
[PASTE YOUR CODE AND THE ERROR HERE]
Profile a slow program and validate speedups against benchmarks. The linked tutorial covers prompts for refactoring legacy code with ChatGPT — restructuring for readability and modernising outdated syntax without changing what the code does.
What you'll take awayRefactoring prompts that clean up messy code while preserving its behaviour.
Starter prompt — paste into ChatGPT
I'm a Computer Science student at HWS. My task: Optimize performance — Profile a slow program and validate speedups against benchmarks.
Tell me what's wrong, explain why it happens, then give the corrected version. Walk through the fixed code line by line with a concrete input.
[PASTE YOUR CODE AND THE ERROR HERE]