Internship to Full-Time to Freelance — The Money Numbers

Developers write about money in percentages and vibes. Here are my actual numbers, stage by stage, and then the analysis that makes them mean something.

Stage 1: college side gigs, second year onward

A few thousand rupees a project. Small sites, small fixes, work I wouldn't put in a portfolio now.

What it taught: someone will pay you before you feel ready, and finishing on a promised day is the whole product. Nobody who paid me then was buying my code. They were buying the fact that it would exist by Friday.

Earnings-wise it was noise. As education it was the highest-return period of my life.

Stage 2: Travelfika, part-time founding engineer, Aug 2022 to Sep 2023

About ₹15,000 a month, roughly $180. Part-time, remote, from a dorm room, while attending college.

For a student that money was meaningful. Measured against the work, it was nothing close to market. I was carrying schema design, backend, and production incidents at midnight for what a US company would pay a junior in a day and a half.

I don't resent it. The title, the ownership, and the reference were worth more than the difference. But I want to be accurate: it was an apprenticeship with a stipend attached.

Stage 3: three internships, 2024

Bosch (Jan–Apr), Sirma (May–Jun), GyanGrove (Jul–Aug). Stipends in the ₹20,000–30,000/month range, so about $240–360.

A 50–100% jump on Travelfika for less responsibility, which tells you how badly early startup work pays relative to structured programmes. I've written separately about what each of those three taught me, and none of it was about the stipend.

Stage 4: Toolpioneers, SDE I, ~Sep 2024 to Mar 2026

Started around ₹9 LPA. Ended around ₹12 LPA. Take-home roughly ₹75,000/month, call it $900.

People find this number surprising in both directions. Indian readers know it's a perfectly normal SDE I salary in Hyderabad. Western readers see $900/month and assume poverty, which it isn't, because ₹75,000 in Hyderabad buys a life $900 in San Francisco does not.

It was also stable in a way I underrated at the time. Same date, same amount, every month, plus provident fund, health cover, paid leave, and a laptop I didn't buy.

Stage 5: freelance, March 2026 onward

Rates $45–65/hour. Typical project $4,000–12,000. First months averaging roughly $3,500–5,000/month gross.

That last figure is an average of numbers that don't look like it. One month cleared close to $6,500. Another came in near $2,200. Same effort, same skill, wildly different output, because a client's procurement cycle doesn't care about my calendar.

So: about four to five times my old take-home, on paper. Now the part that matters.

Gross is not net, and the gap is bigger than you think

When I earned ₹75,000, ₹75,000 arrived in my account and was mine. Freelance gross isn't that.

Off the top of every payment I hold back:

• 30% for tax, set aside the day the money lands, in a separate account I treat as untouchable. Not quarterly, not at year end. The day it lands.
• 10% into a buffer for dead months. Not savings. Payroll for a version of me who has no work in September.
• Health cover, formerly my employer's line item, now mine.
• Equipment and software. Laptop, monitor, phone, SIMs, subscriptions.
• No paid leave. A week off isn't a week of reduced pay. It's a week of zero.
• No provident fund. Nobody is contributing to my retirement but me.

Round it off and roughly 40% of gross is spoken for before I decide anything. A $4,200 month is closer to $2,500 spendable. Comfortably ahead of $900, but not the four-times headline.

The thing that actually did the work: currency

Here's the honest mechanism, and it isn't the rate.

I earn in dollars and spend in Da Nang, Bangkok, Chiang Mai, or Hyderabad, where a good month costs $900–1,200 all in. My income is priced in the client's economy and my costs are priced in mine.

That gap does more for my finances than the rate increase does. If I earned the same dollars and lived in London, I'd be worse off than I was as an SDE I in Hyderabad. Every freelancer selling you a rate story is usually telling a currency story underneath it.

The arbitrage is a position, not an achievement, and it can close. Rupee moves, visa rules change, Vietnam gets more expensive every year. I treat it as a runway I'm standing on, not as permanent income.

Variance is the real cost, not the average

If you take one thing from this post: a good month and a bad month differ by roughly 3×. Mine have ranged from about $2,200 to about $6,500 in the same quarter.

The average is fine. The variance is what makes freelancing hard, because you cannot spend an average, and because a bad month arrives with a psychological tax a spreadsheet doesn't capture. You start thinking about rates when you should be thinking about pipeline.

What I do about it:

I pay myself a fixed monthly salary from the buffer account. Same number every month regardless of what came in. Good months feed it, bad months draw from it. This one habit removed most of the emotional swing.

I keep at least one retainer. Even a small monthly commitment changes a bad month from zero to something.

I hold six months of expenses, achievable at Southeast Asian costs in a way it wouldn't be elsewhere. That number is what lets me decline work priced below my floor instead of taking it out of fear.

I don't let the pipeline empty during good months. The instinct when you're busy is to stop selling, and that instinct is exactly what creates the empty month eight weeks later.

The salary years weren't a detour

It would be easy to frame this as escaping employment. It's the opposite.

The $900/month years are the reason the $55/hour is possible. Bosch is on my site. Toolpioneers is on my site. A year and a half shipping AWS and Retool work for real clients is what makes a founder comfortable wiring a stranger $6,000. I wasn't underpaid then, I was buying credibility at a discount, and I've been spending it ever since.

The number I'm proudest of isn't the rate. It's that I can now turn down a project without doing arithmetic first. That took a salary, three internships, and six months of uneven income to buy, and it's the only financial metric I've found that actually changes how the work feels.