What Mechanical Engineers Actually Do All Day (Hint: It's Not Just CAD)

May 10, 2026

What Mechanical Engineers Actually Do All Day (Hint: It's Not Just CAD)

You told your family you're a mechanical engineer and they pictured you welding rockets. Your roommate pictured you in SolidWorks all day with two monitors and a quiet jazz playlist. Reality is you spent forty minutes this morning explaining to a buyer in procurement why the $4 fastener is not, in fact, interchangeable with the $0.40 fastener, and now you're being asked to justify the cost variance in a meeting at 2:30. Welcome to the job. Here's what the title actually means once the diploma is on the wall.

The Myth of the CAD Monkey

There's a persistent fantasy — held mostly by undergrads and recruiters who've never sat next to an engineer — that the job is 90% CAD. You'll model parts all day. You'll extrude, fillet, mate, render. Maybe you'll get a standing desk and a Spaceball.

The reality is closer to 15-25% CAD if you're lucky and the project's still in early design. The rest of the time CAD is open in a background window while you do everything else. You'll spend more time in Excel than you ever did in your CAD class. You'll spend more time in email than in Excel. And you'll spend more time standing next to a confused machinist holding a print, pointing at a dimension you swore was clear, than anywhere else.

The dirty secret is that experienced engineers actively avoid getting pigeonholed as the CAD person. Once you're the CAD person, you stay the CAD person, and the interesting work — the requirements arguments, the failure analysis, the supplier negotiation — goes to someone else. The new hire who knows every keyboard shortcut in NX is also the new hire who'll be drawing assembly views for the next three years.

What Mechanical Engineers Do When Nobody's Watching

Strip away the org chart and the project management theater, and the day-to-day job of a mechanical engineer is really about answering one question, over and over: will this work, and how do you know?

That breaks down into a lot of small, unglamorous activities:

**Reading.** Datasheets. Material specs. ASME standards your company paid four figures to subscribe to. Old prints from 1987 that someone scanned crooked. The email thread where the previous engineer explained why this gasket is the way it is, which you'll find six hours after you tried to change it.
**Calculating.** Sometimes in MATLAB. Often in Excel. Occasionally on a napkin, then re-done in Excel because nobody trusts the napkin. The classic moment is realizing the answer you need is one equation you remember from sophomore year and a value you have to call a vendor to get.
**Asking.** The machinist, the technician, the supplier rep, the field service guy who's been with the company since Reagan. Half of mechanical engineering is knowing who to ask, and the other half is asking in a way that doesn't make them defensive.
**Documenting.** Because if it's not documented, it didn't happen, and you'll be the one explaining it to the auditor in eighteen months.
**Sitting in meetings.** A lot of meetings. Some of them are useful. The useful ones are the ones where a decision actually gets made.

Notice what's not in that list: rendering pretty assemblies. That happens, but it's a side effect, not the goal.

A Realistic Tuesday

Let's get specific. Here's a composite Tuesday from the life of an early-career ME at a mid-sized manufacturer — say, a company that builds industrial pumps or HVAC equipment or something with both rotating parts and a sales team.

7:45 AM. Coffee. Open laptop. Forty-three unread emails. Three of them matter. One is from the test lab saying yesterday's prototype failed at 60% of the rated cycle count.

8:15 AM. Walk down to the lab. Look at the failed part. The fracture surface tells you it's fatigue, initiated at a fillet that was supposed to be R0.250 and is closer to R0.060 because someone at the supplier substituted a smaller tool. The drawing called it out. The supplier ignored it. Welcome to mechanical engineering.

9:00 AM. Standup meeting. You give a sixty-second summary of the failure. The program manager asks if it'll delay the milestone. You say "depends on the supplier." She asks for a date. You give her one you're 40% sure about.

9:30 - 11:30. Email the supplier with photographs. Call the supplier when they don't respond to email. Get put on hold. Open SolidWorks while you wait and start a redesign that adds a relief groove so the fillet isn't load-bearing in the first place — because fixing the supplier is a six-month problem and fixing the part is a six-hour one.

11:30. Lunch. You eat at your desk because you said you would. You watch a YouTube video about a Swiss CNC lathe and feel slightly better.

12:00 - 3:00. Hand calcs on the new geometry. FEA to confirm. Realize the FEA disagrees with the hand calc by 30%. Spend forty minutes figuring out the mesh was too coarse at the fillet. Re-run. Now they agree within 8%, which is close enough for a coffee meeting but not for a release.

3:00. The meeting at 2:30 that started at 3:00. Justify the $4 fastener. Win the argument. Procurement is unhappy.

4:00 - 5:30. Write up the failure analysis in the format that legal wants. Don't speculate. Stick to what the fracture surface tells you. Submit the new part for drawing review.

5:35. Realize you didn't open CAD for any creative reason today. Go home.

This is what mechanical engineers do. Multiply it by 250 working days a year and you have a career.

The Specialties Nobody Mentions in Sophomore Year

When people ask "what do mechanical engineers do," they usually want a clean answer like "design machines." The honest answer is that the field fragmented into about thirty sub-disciplines a long time ago and most working MEs only really touch two or three of them. A short list of jobs that all wear the "mechanical engineer" badge:

**Design engineer.** What everyone pictures. Owns geometry, drawings, and tolerances. Lives in CAD and PLM.
**Manufacturing engineer.** Owns the process, not the part. Cares less about what it is and more about how it gets made, fixtured, inspected.
**Test engineer.** Breaks things on purpose. Owns the data. Often the only person in the room who actually knows what the product can take.
**Field engineer.** Lives in a rental car. Knows the install crew by first name. Owns the "why doesn't it work in Wyoming" problem.
**Applications engineer.** Translates between customers and the design team. Half engineer, half therapist.
**Reliability engineer.** Statistics, Weibull plots, MTBF arguments. The only ME who actually used the probability class.
**HVAC, structural, fluids, controls, thermal, NVH...** each its own world with its own software and its own arguments.

If you're a student reading this and you don't know which of these you'll end up doing, that's normal. Most working engineers stumbled into theirs.

Q&A: Things Early-Career MEs Actually Ask

Q: Do I really need to learn machining if I'm a design engineer?

You don't need to. You'll just be worse at your job than the engineer who did. There's a particular kind of drawing only produced by people who've never held a part in a vise, and machinists can spot it in three seconds. Spend a week in a shop if you can. Even a community college night class. It pays back forever.

Q: How much math do mechanical engineers actually use?

Less than your professors implied, more than your liberal arts friends believe. The integral you couldn't solve in junior year is now a built-in function in three different software packages. What you do use constantly is dimensional analysis, unit conversion, and the sanity-check arithmetic that catches the supplier who quoted in millimeters and shipped in inches.

Q: Will I ever stop feeling like I'm faking it?

Around year four or five, you'll be in a meeting and realize you're the one being asked the question, not the one asking it, and you'll have an answer. The imposter feeling doesn't fully leave. It just gets quieter and more useful.

Q: Is it worth getting a PE license?

Depends entirely on industry. HVAC, civil-adjacent, consulting, anything with public safety stamps: yes. Product design at a manufacturer: usually no, but it doesn't hurt. Ask the senior engineers in your group. If none of them have it, you probably don't need it.

The Identity Question

There's something the job description doesn't tell you, which is that being a mechanical engineer is also a social identity. You'll catch yourself measuring things you don't need to measure. You'll have opinions about which brand of caliper is acceptable. You'll explain torque specs to family members who did not ask. You'll see a guardrail bolted to a highway overpass and instinctively check the fastener pattern.

This is fine. It's even the point. The job rewires how you look at the built world, and after a few years you can't unsee it. The bridge becomes a free-body diagram. The dishwasher becomes a thermal-fluid loop with a control system. The office chair becomes a four-bar linkage you have opinions about.

Whether that's a curse or a feature depends mostly on whether you went into the right field. If you find yourself sketching mechanisms on cocktail napkins, you probably did.