From nine to six I am a senior mobile engineer on a two person team, shipping features that move real money for real users. From nine to midnight, some nights, I am a one person software company with zero meetings and a customer base I can count on my fingers.
Two of those night products are live right now: CQL, a native macOS database client with a built in VPN profile manager, and AIMeter, a menu bar app that tracks quota across thirteen AI providers. Neither started as a business. Both started as an itch I had at my day job that no existing tool would scratch.
People ask how the two lives fit together. Honestly, the more interesting question is what each one fixes about the other.
The day job teaches consequences
At work, code has weight. A bad release reaches thousands of people before lunch. So the day job drilled habits into me that no side project ever would have: review discipline, migration plans, the paranoia of asking what happens when this fails at 2am. You do not learn that from a tutorial. You learn it from being the person who gets the message when it breaks.
Every one of those habits made my products better. CQL handles credentials the paranoid way because I have seen what casual credential handling costs. AIMeter treats every provider response as hostile until proven otherwise because production taught me that a 200 status code is an opinion, not a fact.
The products teach ownership
At a company, even a small team, you own a slice. On my own products I own the pricing page, the icon, the release notes, the refund email, the bug that only happens on one user’s machine. There is nobody to escalate to. I am the escalation.
That changes how you think at the day job too. Once you have personally answered a confused customer email caused by your own UX decision, you stop treating design reviews as someone else’s meeting. Ownership is a muscle, and side products are the gym.
The rules that keep it sane
Hard separation. Work code never touches personal machines, personal products never get mentioned in work repos. Not one import, not one commit message. The moment those streams cross, both sides get complicated in ways that are never worth it.
Nights are optional. The day job is a commitment. The products are a choice, every single evening. Some weeks the choice is sleep, and the products wait. That asymmetry is the entire reason this is sustainable.
Ship small, ship real. A night product that stays at ninety percent forever is a hobby that generates guilt. Cutting scope until a thing can actually go live taught me more about product thinking than any book.
One user is enough. Both products started with a user base of exactly me. Me is a demanding customer. If it survives me, it can meet strangers.
Why bother
Not the money, at least not yet. The honest answer: the products are where I still get the feeling that got me into this field. A whole thing, built end to end, by hand, that did not exist before and now does. The day job pays for the life. The night work reminds me why I picked this life in the first place.
If you are an engineer with an itch and a free evening, my advice is unoriginal and correct: build the small tool you keep wishing existed. Worst case, you learn something your job would never teach you. Best case, a stranger pays for it, and there is no feeling in this industry quite like the first time that happens.