All internships
Schoolabe · Work from home
AI / ML Intern

- Stipend
- Paid · amount is in the offer letter
- Duration
- 3 months
- Commitment
- Part-time · about 20 hours a week
- Profile
- AI / ML
What you will do
You are placed on a project that needs a small model or an AI feature, not on one shared lab. The data and the question come with that project.
The work stays small enough to run on a laptop. Fit a simple model or wire a narrow feature, report the result from a run you did, and show one case it gets wrong. You explain that case in plain language. You do not train a large model, and you do not invent a benchmark score.
At about 20 hours a week, a normal week is one of those pieces, including the revision after someone re-runs your code.
What you hand in
- The notebook or the change
- The data the project gave you, one simple model or feature, and the score or output from a run you did.
- One mistake
- A case the model got wrong, and a sentence on why that mistake is plausible from the inputs.
- The limit
- What this model is for, and what it is not for. A student should not leave thinking the score is a product metric.
Skills
- Python
- You can load a table and fit a model in a notebook. scikit-learn is enough. A deep-learning stack is not required.
- A small model
- A classifier or a regression on a few hundred rows. You can say what it predicts.
- Error cases
- You look at a wrong prediction instead of only quoting the score.
Who can apply
- You have trained one small model on a public table and can show the code and the score.
- You can explain a wrong prediction without hiding behind the metric.
- About 20 hours a week, for 3 months, from home.
What to send
- The application form on this site. It is free, and you need a Schoolabe account.
- A notebook on a public dataset: the model, the score from your run, and one mistake.
- Schoolabe writes back if they want to continue. A stipend amount is in the offer letter, and only if you are accepted.
Free to apply. You need a Schoolabe account. Selection is not guaranteed.