Is it okay to use ChatGPT for a take-home assignment in a grad data science interview?
Zhang Xinyi 张欣怡@xinyi_in_melb
Hi everyone, I am finishing Master of Data Science and applying for grad roles. I got a take-home assignment from one company: analyse a customer churn dataset, build a model, write short report, 5 days.
The instructions don't say anything about AI tools. I have question:
- Is it okay to use ChatGPT or Copilot to help? For example for pandas syntax, or to check the English in my report (my English is not native).
- Should I say in the report that I used it? I am worried they will think I can't do it myself.
- If they ask me in the follow-up interview to explain my code, and some code is from AI, how to do?
My visa time is not so long, so I really want to do this one right. Thanks in advance!! 🙏
Accepted answer
hi xinyi! my honest view as someone who reviews take-homes at my bank:
1. using it for syntax and checking your English is fine lah. everyone googles pandas syntax, AI is the new google for that.
2. yes, mention it, one line at the end: 'I used ChatGPT for syntax lookups and proofreading. All analysis decisions and code are my own and I can explain them.' this reads as honest and mature, not weak.
3. this one is the important one. don't submit any line you can't explain. before you submit, go through every cell and say out loud why it is there. if you can't, rewrite it your way.
what we actually look for: did you check for leakage, did you handle the class imbalance (churn is always imbalanced), did you explain the business meaning. AI won't think about those for you. you can do this! 💪
Get the best of #my-model-is-fine weekly
One email each Monday with the most useful posts. No spam; unsubscribe in one click.
Comments
Sign in to join the conversation.
Sign in to commentAccepted answer
Nurul Aisyah@nurul_sql
hi xinyi! my honest view as someone who reviews take-homes at my bank:
1. using it for syntax and checking your English is fine lah. everyone googles pandas syntax, AI is the new google for that.
2. yes, mention it, one line at the end: 'I used ChatGPT for syntax lookups and proofreading. All analysis decisions and code are my own and I can explain them.' this reads as honest and mature, not weak.
3. this one is the important one. don't submit any line you can't explain. before you submit, go through every cell and say out loud why it is there. if you can't, rewrite it your way.
what we actually look for: did you check for leakage, did you handle the class imbalance (churn is always imbalanced), did you explain the business meaning. AI won't think about those for you. you can do this! 💪
Zhang Xinyi 张欣怡@xinyi_in_melb
Thank you so much Nurul!! The one line disclaimer is very helpful, I will use it. And I will check the imbalance, churn rate is 14% in the data 🙏
Mo Siddiqui@mo_boston
agree with nurul 100%. also the report matters more than the model. a logistic regression with a clear explanation of the top 3 drivers beats an xgboost nobody can explain. hiring managers read the first page and skim the rest lol
Dana Kowalski@dana_ml
Everything above is right. One addition from someone who has designed these assignments: compare against a trivial baseline and say so in the report. 'Predicting that nobody churns gives 86% accuracy; my model gets 0.61 recall on churners at 0.45 precision' tells the reviewer you understand the problem. Very few candidates do this, and it takes ten minutes.
Zhang Xinyi 张欣怡@xinyi_in_melb
Oh, the baseline sentence is very smart, I didn't think of it. Thank you Dana! I read your post about evals, now I understand it better with this example
Kenji Tanabe@kenjitanabe
On question 3: if they ask about code you can't fully explain, say so, and explain how you verified it. 'I used a suggestion here and checked the output against a manual calculation' is a good answer. Pretending is the only bad answer.
Bruna Schmitt@bruna_dados
And look for a date column! If the churn data has dates, split by time, not random. Most candidates don't, and reviewers notice. Boa sorte, Xinyi! 🧉