[throwawaycash]

I make about 0-300k/year fixing apps that others either weren't able to finish or where they architected themselves into scaling issues. Aside from that, I am a part time programming consultant who makes about 80k a year (I have a chronic illness so I spend a lot of time not working, and it's helpful for me to work in intense bursts with rests in between. It's difficult for me to have normal full time employment in the US because of the chronic illness.)

It averages out to about $250k/year on 30/h/week

[neuronsguy]

This sounds cool. Two questions:

1. Without giving away secrets, how do you find such clients?

2. You must have some great stories. Would you care to share any?

[throwawaycash]

1. I found my first few clients literally through happenstance -- people were describing problems they were having with enough clarity that my 20 years of dev experience made it fairly clear that the problems were recoverable. Now, I mostly find customers through referrals and by attending networking events and talking to people who sound like they have both money and problems.

2. My customers have the expectation of confidentiality. I will make these generalizable statements though: understanding ORMs and understanding SQL are related but not identical; CAP is hard; picking keys is hard; confounding your data model and your interface is painful for users; sharding is hard; pick UX libraries based on how consistent they are not on how new they are; developers that are producing 10x the code may not be producing 10x the value; if week on week, your project is consistently n/n+1 * 100% done, you have a estimation problem and just the fact that you're agile doesn't help with this; the whole point of tags was that not everyone agrees on categories; queuing theory is real; your spreadsheets aren't as consistent as you think; the fact that an engineer can figure out the bigO of a sort doesn't mean they understand efficiency in real life.

[throwawaycash]

(2 cont) your PM may only be paying attention to the user interface and not what lies beneath it; changing plans all the time isn't agile, it's just chaos; you have to keep an eye on whether you're doing TDD or just architecting mocks; n+1 queries are hard once you start encountering real data sizes.

[srednalfden]

For someone with health issues that can't work full time (or has to work freelance) what tech skills would you recommend?

[throwawaycash]

I recommend API development and not front end development. I recommend companies that don't have on-call requirements for engineers. I recommend against directly working with a company that has agile processes, if you can manage it.

You'll be fine if you can produce, but if you can only produce on the week level and not the day level, you'll have troubles with companies that want to talk about the progress you're making day to day.