On the first day of an internship, plenty is usually ready and waiting: a desk, a second monitor, a license and a tour past the coffee machine. What is rarely ready is the one thing that actually decides how the next ten weeks will go: a list of well-defined tasks, a set of permissions that fits a beginner and a working method that explains how things are done around here.
Because the same intern who, with the right preparation, leaves behind a permanently cleaner dataset will quietly wreck your folder structure without it. Getting an intern or junior engineer productive in CAD is therefore mostly a matter of thinking ahead: about what they will do, what they are allowed to touch, and where their work has to pass before it counts. In this article you will read which tasks genuinely pay off, which permissions you grant and which you withhold, and how templates and fixed working methods, if you like supported by tooling from a toolbox such as Thundercad, shrink the gap between junior and senior work.
To be clear: this piece is about interns and starting engineers who join temporarily or under close supervision. The complete onboarding track for a permanent new colleague is a story of its own; we covered that earlier in Onboarding a new engineer: from three months down to three weeks.
Three requirements for every intern task
The difference between an intern who adds value and an intern who causes damage rarely lies in the person. It lies in the work you hand them. A good intern task meets three requirements:
- Well-defined. There is a list with a beginning and an end: these fifty models, this one completed project, this folder. "Just tidy up the server a bit" is not a task, it is a trap.
- Checkable. The result can be verified without redoing all the work: a spot check, a before-and-after comparison, a review per drawing.
- Reversible. If it goes wrong, it can be rolled back. The intern works on a copy, in a workspace of their own or in an environment with version control, and never directly in released data.
Hold every candidate task against this yardstick and most discussions about what an intern can handle settle themselves. Drawing work on a rush order that has to ship this afternoon: not well-defined, barely reversible, so no. Giving fifty old parts a proper description: three ticks, so yes.
Tasks that genuinely pay off
Overdue cleanup work
Every drawing office has a pile of jobs nobody ever finds time for: models full of leftover sketches and dead features, folders with stray export files, projects that were never properly closed out. For a senior that work is mind-numbing; for an intern it is the ideal introduction to the dataset and the way of working. With a tool like Model Cleaner, which clears leftover geometry and clutter out of models, even a beginner can do this safely: the tooling decides what can go, the intern works through the list and the senior spot-checks each project.
Updating metadata
Missing descriptions, empty material fields, item numbers that were never filled in: metadata backlog exists in nearly every company and costs search time every single day. This is junior work par excellence, provided the fields are fixed. The iProperty Panel in Thundercad provides exactly that: a configurable data card per document type, so the intern sees precisely which fields need filling and cannot skip one or invent their own.
Drawing work under review
Production drawings of existing, simple parts are perfectly good intern work, on one condition: nothing goes out the door without a second pair of eyes. Have the intern work from an example drawing and treat every sheet as work to be reviewed. How to keep that check light and workable, even in a small team, is something you can read in Four eyes on every drawing, even when there are only two of you.
Permissions: what you grant and what you withhold
Handing out permissions to an intern is not a matter of trust but of setup. Your best engineer makes mistakes too; the difference is that an intern does not recognize them. So arrange the environment in such a way that any mistake stays small by definition.
What you do grant: read access to the projects they will be working with, write access in a workspace or practice project of their own, and a personal project file that points to the same libraries and templates as the rest of the team. What you do not grant: write access to libraries, templates and standard parts, the right to release drawings, and in Vault the rights to delete files or change lifecycle states.
If you work without Vault, the file server does the same job: library and template folders read-only, live order folders too where possible, and one folder of their own where the intern is free to build. If you work with Vault, give them a role of their own with permissions of their own, not the account of the previous intern with everything that has stuck to it over time.
Cleanup and metadata work only become truly intern-proof with tooling that sets boundaries: Model Cleaner decides what can safely go, the iProperty Panel guards which fields get filled.
Try 30 days freeTemplates and fixed methods shrink the gap
Why is a senior faster? Rarely because they click faster. They know where everything lives, what things are called and what the shop floor expects on a drawing. That is exactly the difference you can shrink with fixed working methods: the tighter the frame, the smaller the gap between junior and senior work.
In practice: a drawing template in which the title block fills itself from the iProperties, a fixed naming convention for files, an example project that shows what a clean assembly looks like, and a short checklist per drawing. Within a frame like that, a junior delivers work that at first glance is hard to tell apart from senior work. Without a frame they have to guess every convention, and guessing turns into correcting.
As an assumption for the math: reviewing a drawing made to template and checklist takes about five minutes. The same review of a sheet where naming, title block and layout also need straightening out quickly runs toward a quarter of an hour. Fixed methods pay off twice: the intern works more independently and supervision takes less time.
Supervision in fifteen minutes a day
Supervision fails in two ways: too little, after which the intern flounders, dares not ask anything and delivers something unusable after six weeks, or too much, after which the senior spends half days sitting next to them and the internship costs capacity on balance. The workable middle is a fixed, short rhythm.
A fixed fifteen minutes at the start of the day is enough: go through yesterday's work, answer questions, define today's task. Also have the intern keep a simple logbook: what was done, what stood out, which questions came up. That logbook is worth gold at the end of the internship, because everything an intern stumbles over is implicit knowledge the next new colleague would otherwise have to guess at too.
Deliberately plan the final week as a wrap-up week: finish the work, hand over the results, and set aside half an hour in which the intern tells you what struck them. That list is often more honest than many an internal improvement plan.
Frequently asked questions
What task do I give an intern on day one?
Something small with visible results and zero risk: checking and completing the metadata of one finished project, for example. The intern gets to know the dataset and the naming conventions, you immediately see how they work, and nothing can break.
Should an intern get access to Vault?
Yes, but with a restricted role of their own: read and add where needed, no release rights, no delete rights, no state changes. That way they learn to work the way your team works, without one wrong click ending up in released data.
How do I keep supervision from costing more than the internship delivers?
Pick tasks that are well-defined, checkable and reversible, and provide tooling that sets boundaries instead of freedom. You can try the cleanup and metadata tools in Thundercad free for a month: plenty to cover a complete internship.