Spend your hours on signal, not code
With a day job, your time is the constraint, so spend it where it returns the most. Reading forums, subreddits, and review sites for the same complaint repeating is something you can do in twenty-minute gaps. Customer conversations are the other high-value activity. Building a product is the lowest-value thing you can do this early, and it is also the biggest time sink, so do not start.
Use async research to fit your schedule
A lot of validation does not need a calendar invite. Posting in a relevant community, reading competitor reviews to find gaps, and sending cold outreach all work on your time, not the buyer's. Batch these on a weekend, then let responses trickle in during the week. Async outreach is how full-time founders get real customer input without taking calls during work hours.
Schedule the calls you cannot avoid
You will still want a few live conversations, because tone and hesitation tell you things a survey cannot. Book them for early mornings, lunch breaks, or weekends, and keep them to twenty minutes. Five to ten focused conversations is enough for a first read. Ask about their past behavior and real spending, not their opinion of your idea, so a short call still gives you honest signal.
Protect yourself and set a real deadline
Check your employment contract and any IP or non-compete clauses before you put serious time in, and keep this work clearly separate from your employer's equipment and hours. Then give yourself a hard deadline, four weeks is reasonable, so the project does not drift forever in the background. A deadline forces you to test the riskiest assumption first instead of polishing safe details.