Should every model type I have read about go on the page?

No. List methods you have used on data you can describe. A chapter you read is not a skill line.

What if the model never shipped?

Say that. An analysis a lead used, or a model that failed a check you can name, is more honest than a silent implication that it ran in production.

Data scientist resume tips

A data scientist resume is a record of questions you framed and the analysis or model someone could use. The library is supporting detail. A reviewer still needs the decision, the table you actually had, and whether the work stayed in a notebook or reached a person who acted on it.

Cover letter for a data scientist

What to put first

Say the domain and the kind of question: ranking, forecasting, classifying a case, or measuring an experiment. Then name the data you were allowed to use and who read the result. Put the closest example to the posting first, and put the title you want in the heading.

Question before the model

Start with the decision the work was meant to inform. Then the features or the table, then the method. A model name with no question reads like a course project, even when the job was real.

What you could see

Say whether you defined the label, inherited a warehouse table, or cleaned an export. Those are different jobs. Do not claim you built the pipeline if you only trained on a file someone else refreshed.

A result you can explain

If you mention quality, say how you judged it and who decided the cutoff. Skip a score you cannot define out loud. A model that was tried and not shipped is still worth a line if you say it was not shipped.

An example line

Adapt this only where it is true for you. It is a shape for a bullet, not a claim that a particular person did this work.

Framed a ranking question for a support lead, trained on tickets they already stored, and wrote the notes they used to decide which cases to review first.

What to leave off

Leave off libraries you cannot use without a tutorial open, metrics you did not define, and a production system you never wired up. Leave off a claim that a model changed the business by a figure you cannot walk through.

Questions

Should every model type I have read about go on the page?: No. List methods you have used on data you can describe. A chapter you read is not a skill line.
What if the model never shipped?: Say that. An analysis a lead used, or a model that failed a check you can name, is more honest than a silent implication that it ran in production.