Civil engineer to UX designer, and what it actually cost

I spent two years detailing steel reinforcement, resigned in July 2024 with nothing lined up, and started as a design intern nine months later on 15 April 2025. Sixteen months after that I was leading design. This is the version with the dates in it.

Most accounts of this switch are written by people selling a course. They skip the gap, round the timeline up, and never mention the salary. I want to put the specifics somewhere, because when I was deciding whether to do it, specifics were the only thing I actually wanted and the only thing nobody published.

What the engineering job actually was

Not buildings. Rods. I was a rebar detailer, which means the drawings I produced said where every piece of steel reinforcement goes inside a concrete structure: diameter, spacing, lap length, bend angle. A detailer’s entire world is whether the bars clash at the column junction, because that is where four directions of steel try to occupy the same 300mm of space.

It is more demanding than it sounds, and the people who are good at it are genuinely good at it. It is also the most unforgiving brief you will ever receive. Your drawing gets poured in concrete by someone who will not read your notes, cannot ring you, and is not going to stop and think about what you probably meant. If it is ambiguous, it is already wrong. I think about that every time I hand off a Figma file now.

I was good at it. That was the problem. Getting better meant getting more precise at the same 12mm bar for another four years before anyone would let me near a real structural decision. The ladder was completely visible from where I stood, I could see roughly how long each rung took, and I did not want any of them. There is a shorter version of that in why I quit rebar detailing.

How design turned up, honestly

It was a semester break in college. My laptop was mostly running AutoCAD. I was bored, the Google UX Design course on Coursera cost nothing, and I started it for no better reason than that. There are seven courses in that certificate. I got through a few, term started again, and I stopped.

Then I got placed, and for about a year I did not think about it at all. That is the part that gets left out of these stories. The interest did not build steadily into a decision. It went quiet, completely, for a year, while I did the job I had trained for.

What brought me back was not the course. It was noticing that the thing I actually liked about engineering, that a design is either buildable or it is not, existed in product work too. Same constraint, same test. Without the four year wait to touch anything real. I wrote about that in more detail in how I found design on a free Google course.

Quitting

July 2024. I had just finished a set of detailing drawings that were, by every measure the department used, correct. Clash free. Scheduled. Checked. And I realised I could produce that same set every week for four more years and the only thing that would change was how fast.

I resigned with no portfolio, no network in design, and no plan beyond a bootcamp I had not paid for yet. That was reckless and it worked, and those two facts do not cancel each other out. I would not recommend it to anyone. I also would not undo it. The longer version is in the day I quit.

The nine months nobody posts

Between the resignation and the first pay cheque for design work there were nine months. They broke down roughly like this.

Two months quietly certain I had made a mistake. Five months at DesignBoat in Bengaluru, rebuilding one fintech project called Grownest until it stopped being ugly. Two months applying with a portfolio that was mostly nice looking screens with no reasoning underneath any of them, which is the single most common failure I now see in other people’s portfolios and I recognise it because it was mine.

I started on 15 April 2025. As an intern. After an engineering degree and two years of professional work. Taking the internship was the right call and I will argue with anyone about it: it was the shortest route to real projects with real constraints, and nothing in the previous nine months had given me those. More on the gap in nine months from quitting to a design job.

What actually transferred

This is the part I was most wrong about. I assumed the engineering was sunk cost. It was not, and the transfer was not the vague thing people mean by transferable skills.

Design systems are bar bending schedules. In rebar detailing you do not draw every bar. You define types: B1, B2, B3. Each has a shape code, a diameter, a length. The schedule then says how many of each go where. Change B2 once and every floor updates. The first time someone explained components and variants to me I thought, we already do this, we just do it in AutoCAD and nobody calls it a system. The difference is the cost of being wrong: a wrong schedule costs lakhs and three weeks on site, a wrong component costs a developer twenty minutes and some goodwill. Both are worth getting right, but only one lets you fix it afterwards. That comparison is the whole of this post, if you want it on its own.

Civil drawings carry a tolerance and UI specs do not. A structural drawing states how wrong the build is allowed to be before it counts as a defect. Cover to reinforcement, plus or minus 5mm. Slab level, plus or minus 10mm. Nobody expects site to hit the drawing exactly, so the drawing says what site is allowed to miss by. Then I moved into product and watched teams ship specs with no tolerance at all: one number, one perfect screen, one width, one language, a name that fits. Then the name is Venkataraghavan, the connection is 3G, and the string is German. Edge cases are not edge cases. They are the tolerance, and they belong in the spec.

Labelling is not tidiness. On site, an unlabelled element becomes a query, and a query becomes a phone call at 7am from someone standing next to a concrete pour that cannot wait for you to wake up. Figma has no concrete pour. It just has a developer quietly deciding what “Frame 247” meant, and you find out which decision they made in review. That habit is the subject of why I name Figma layers properly, and it is the single cheapest thing I brought across.

Intern to lead in sixteen months

April 2025 to August 2026. I know how that reads. It reads like a small company where somebody left, and both of those are true. It is worth being precise about what a fast title at a small studio buys you and what it does not.

What it bought: three projects end to end, two more with a team of two, nobody senior checking my work, and every bad decision landing on me within a fortnight instead of a year. What it did not buy: anyone who could tell me the work was wrong before the client did. I am still filling that gap deliberately, and it is a large part of what I am looking for next.

If you are weighing a small studio against a structured team, that is the actual trade. The studio hands you the reps. It does not hand you the standard. You have to go and find that yourself, and most people underestimate how much slower that is. The full accounting is in intern to design lead in sixteen months.

The other thing a fast title does is remove your excuses. When there is nobody above you, a weak decision does not get caught in review, it gets caught by a client three weeks later in front of other people. That is a harsher feedback loop than a design critique and it is a much faster one, which is most of why sixteen months was enough to change how I work rather than just what I was called.

The work that taught me most

Wedoura, a two sided wedding marketplace, is where I learned that research without a budget is still research. There was no user base to interview, so I signed up as a test user on established Indian wedding marketplaces and read public vendor complaints. One pattern came out of it: vendors who listed a WhatsApp number answered reliably on WhatsApp and almost never through the platform’s own inbox. They were not ignoring buyers. They were filtering by channel before reading a word. That reframed the brief from get vendors to reply into make an in platform enquiry worth as much as a direct message.

GoWheels, a four role mobility platform, is where handoff taught me something a course could not. I had missing empty states, undesigned error states, and data appearing on one interface with no clear source anywhere else in the system. My developer walked me through the gaps. Designing a connected system means accounting for every piece of data across every surface it touches, not just the screens where it is most visible. Screens are the output of a system. The system has to be designed first.

Both case studies are honest about what shipped versus what I would redesign now, which matters more to me than either of them looking finished.

If you are considering this

Four things I would tell the version of me deciding in 2024.

Do not quit without a plan. I did, it worked, and I still think the advice stands. The nine months were survivable because of circumstances that were partly luck, and luck is a poor thing to build a plan on.

Take the internship. A title is a rounding error next to the number of real projects you touch in your first year. Mine was worth more than another six months of polishing a portfolio nobody had asked to see.

Your old field is the differentiator, not the liability. There are a lot of junior product designers. There are very few who can explain a design system by comparing it to a bar bending schedule, and that explanation has opened more conversations for me than any screen in my portfolio.

Write down the reasoning, not just the outcome. The portfolio that got me hired was the one where every decision had a rejected alternative attached to it. The one before that was prettier and got nothing.

Expect the timeline to be lumpy rather than steady. Mine reads as nine months in a sentence, but it was two months of nothing useful, five months of concentrated work, and two months of applying while wondering whether the five had counted. Anyone presenting this as a smooth ramp is either misremembering or selling something. Budget for the flat parts, because they are where most people stop.

Where I am now

Lead UI/UX designer, based in Chennai, open to relocating. I work on systems with multiple roles, real operational constraints and edge cases that matter, which is probably not a coincidence given where I started. If you are hiring for that, or you are somewhere in your own version of the nine months and want to compare notes, my email is on the about page and the rest of the writing is in the rest of what I write.