Home-Assignments Cheat-Sheet
- Anastasia Karavdina
- Jun 23
- 2 min read

Guys, let's talk serious now. I understand that we’re all in a rush, but can we please, please stop blindly using GPT to solve home assignments we’re not sure how to tackle ourselves?
If you get a home assignment, just uploading it to ChatGPT and relying on whatever comes up is very silly (to say the least)!
Here is your step-by-step guide if you're looking for a job in the Data field:
Read the task. Twice.
If you understand it, great, go ahead: draft the solution and start coding. Well, there are a bit more steps than that, but you know everything — just go ahead!
If you don’t understand the task. Like, not at all. Like it has all these “Marketing Mix Modeling” or “building FastAPI” things and you have no clue what any of it means.
You have a choice:
Reply back politely, thanking them for the opportunity and be proud that you saved time for both sides.
Take up the challenge, but be prepared: you most probably won’t get this job. This will be very stressful learning, under a lot of (time) pressure. If you have plenty of time and the field seems to be interesting, why not. But accept from very beginning that probability to get the job is very-very low.
Identify the topic the task is about. Here you can use ChatGPT or another LLM—not to solve the task, but to understand what it is. Is it time-series forecasting, hypothesis testing, or something else?
Google the topic and 𝐬𝐩𝐞𝐧𝐝 𝐚 𝐝𝐚𝐲 𝐫𝐞𝐚𝐝𝐢𝐧𝐠 everything you can find about it. To save time, you can find someone experienced in this topic (e.g., via keywords on mentorcruise.com), get them on a call, and ask their advice about getting the basics of the topic and what one should keep in mind when solving a task like this. Again, your goal is not to find someone to solve it for you, but to learn as much as you can, so you can work out a solution yourself.
If two days have passed and you still have little clue what to do, reply back politely, thank them for the opportunity, and be proud that you still saved time for both sides.
If you’ve managed to grasp the topic, come up with some kind of solution, and are willing to give it a shot - make sure you prepare a decent, logical report with a clear structure. More often than you think, the report is more important than the code you wrote. Seriously!
In summary: it’s totally fine to use LLMs to understand a task and to "deep-research" a topic you haven’t come across before. But at the current stage, it will not give you a good solution for a vaguely formulated home assignment, and the “initial” solution you’ll get from an LLM will most probably mislead you. Almost every week I’m reading someone’s home assignment and believe me, I can always smell if it’s inspired by a GPT hallucination. Avoid this trap.
And good luck with your job search! 🍀
Comments