Written by Victor Smushkevich
When people ask for career advice, they usually ask whether to follow their passion. I would change the question. Passion is a good reason to start something. It is a poor way to decide whether that something should pay your rent. I build a consumer mobile app, and the lessons that have held up are the ones I could defend to a skeptical stranger, not the ones that felt good.
Learn the hard part by building something a stranger will judge
I run Mold Scanner AI, an app that identifies household mold from a photo. It is pre-launch and paid. I am a product builder, not a doctor, industrial hygienist, or inspector. That boundary is a career lesson in itself: knowing what you are not qualified to claim is part of doing the job well.
Photo AI teaches humility quickly. An image model reads color, texture, pattern, and shape. It can tell you a surface resembles something. It cannot see behind drywall, tell you how long a spot has been damp, or correct for a bad photo. A dark stain, a patch of dirt, and a smudge of soot can look alike in poor light and at a poor angle.
So the real product decision is not the model. It is what the app is allowed to say, and how plainly it admits doubt. Designing for trust in an AI answer mostly means putting the limit right next to the answer, so the person knows what to do with it. If you want a job in this field, that judgment is what you are really demonstrating. A model architecture is easy to copy. Restraint is harder.
Distribution and pricing are half the work
Shipping to the App Store is its own craft. Review guidelines, privacy disclosures, screenshots, and onboarding all shape whether anyone gets to the good part. Builders who love the model work tend to leave this for last, and it is where many projects stall.
The on-device versus cloud choice is a similar judgment call. On-device is fast and keeps photos on the phone, which matters when the picture is of someone’s home. Cloud lets you run larger models and improve them without waiting on a release. Neither wins outright. You pick based on what the person using it needs to trust, and you should be able to say why in one sentence.
Pricing before launch is the part people avoid. I have no measured results to hand you, so I will give you a principle instead. A price is a hypothesis about who values the outcome and what it replaces. Deciding early that a product is paid forces you to name that person. A vague plan to sort it out later postpones the hardest question, and it leaves you guessing about whether you built a business or a hobby. If you are weighing tools in this category as a buyer, our guide to the best mold detection app covers what to look for.
When a side project should stay a side project
Not every passion should become a job. Here is how I would sort it.
Keep it a side project if the making is the whole reward, if you cannot name who would pay, or if you suspect income would push you to build things you would resent. A side project has no obligations, and that freedom is worth protecting.
Consider making it the job if a stranger’s problem keeps pulling you back, if you can describe the next hard decision without stalling, and if you enjoy owning the unglamorous parts. On a small team, nobody hides in a specialty. The same person may argue about the model in the morning and rewrite a store listing in the afternoon. If that sounds like a good week, that is real evidence. If it sounds exhausting, that is real evidence too.
Quitting a job is a bit like closing a wall. It is cheap to check first and expensive to reopen. I wrote about that habit for houses in Check moisture before you close the wall, and it transfers cleanly: verify before you make a change hard to undo. Build a small version. Put it in front of strangers. See whether you still want it after the novelty is gone.
One rule to take with you: do not leave a job for a feeling, leave it for a decision you can defend out loud. Who pays, what your product may honestly claim, and what you will do if it stalls. If you can answer those three, your passion has earned a plan.
Author Bio:
Victor Smushkevich is the founder of Mold Scanner AI, a consumer AI app focused on household mold and moisture awareness.
