You don’t need the best engineer you can find. You need someone who will still be there in eighteen months when the product hasn’t taken off yet and the code is embarrassing. Skill is table stakes. Commitment under uncertainty is the actual filter, and it’s much harder to see in a first conversation.

Where to actually find one

• Former coworkers you’ve already shipped something with, even a small internal tool. You already know how they handle a deadline and a disagreement.
• Hackathons and build-in-public communities, where you can watch someone finish something small before you trust them with something large.
• Not cofounder-matching apps as a first resort. They work, but the hit rate is low and the vetting is entirely on you — treat matches there as a lead, not a candidate.

What to test before you commit

Work together on something real for two to four weeks before either of you says the word “cofounder.” A paid contract, a hackathon, a real feature for an existing idea — doesn’t matter, as long as it has a deadline and a disagreement somewhere in it.

Watch for:

Signal	Good sign	Bad sign
Disagreement	Argues the point, then commits to the decision	Goes quiet, then does it their way anyway
Deadline pressure	Communicates delays early	Goes dark until the deadline, then apologizes
Credit	Names your contribution unprompted	Talks about “my” project in front of others
Money talk	Brings it up directly, early	Avoids it, hints, waits for you to raise it

That last one matters more than people expect. Someone who can’t talk about money before there is any money will struggle to talk about it once there’s real money on the table.

The agreement to sign before you write code together

A founder agreement, even a short one, before the second real week of work. It should cover:

• Equity split and vesting. Standard is a four-year vest with a one-year cliff, even between friends. This protects both of you if one leaves early — nobody wants to co-own a company with someone who did one month of work.
• IP assignment. Whoever’s laptop the code was written on doesn’t matter; the company needs to own it, in writing, from day one.
• Decision rights. Who has final say on product, on hiring, on fundraising terms, when you disagree and can’t reach consensus.
• An exit clause. What happens if one of you wants out in year one. Decide this while you like each other — it’s much harder to negotiate after a falling out.

The real test

Money and equity conversations are the actual audition. Anyone can write good code under low stakes. Watch how they handle the conversation where interests might not perfectly align — that’s the relationship you’re actually signing up for.

---