After finishing with my master's in cs last June, I decided I wanted to create my own business.
In many cases, you can test your MVP hypothesis with no-code tools that will be faster and better than whatever you can develop in a week.
Go easy on the domain name. Don't make it too specific if you think you might expand to a market that won't feel represented.
If you need to do outreach to contact your clients, choose a reputable email provider (e.g. Gsuite). Otherwise, you'll experience a higher rejection (bounce off, ending up in spam ...).
Don't reinvent the wheel. Too often I end up with issues on services that aren't really my core business... Specifically, images and videos are a pain.
I used this template from MaterialUI + ReactJS for both sites. That wasn't ideal given that it's a Client-side rending page. I had to refactor the thing to work with SSR for SEO. Speaking of the devil... SEO takes ages to kick in (i.e. a span of months).
I learned tons by developing my own backend (GCP nodeJS GraphQL MongoDB). Still no clue on how that would scale. My intuition is that it would have been better to use a service like Firebase for scale and security.
Distribution is king. Start trying to build an audience asap. Especially so if your target customer is other developers or IHs. In that case, working on boosting your Twitter presence will pay off.
Build in public. Your business gets more exposure and can interact with others like you. It helps with not feeling as lonely or a loser when things don't go right.
Building in public alone won't bring you exposure. Interacting with others and bringing value will (even if not in your target)
I hope it helps someone out there :)
Follow my journey on Twitter