A technical cofounder is not a hire. The person builds the first product, makes the architecture decisions the company lives with for years, and owns a founder-sized share of the stock. The search should be run like a search for a business partner, because that is what it is.

The failure pattern is familiar. A non-technical founder with an idea and a slide deck offers 5% to an engineer, the engineer says no, and the founder concludes that engineers are hard to find. The offer was the problem. Engineers who could start a company on their own are being asked to take employee-level equity for founder-level risk.

Start with what the company needs built

Before looking for anyone, write one page that answers three questions:

1. What does version one have to do? Not the roadmap. The smallest product a first customer would pay for or use weekly.
2. What kind of engineering is it? A mobile consumer app, a data pipeline, a hardware prototype and a B2B web app need different people. “Technical” is not one skill.
3. What has the non-technical founder already done? Customer interviews, a waitlist, letters of intent, a pilot with a named company. This is the evidence a strong engineer will ask for first.

A founder who cannot fill in the third answer is not ready to recruit a technical partner. The better use of the next month is talking to customers.

Where technical cofounders actually come from

There is no single channel. The places below are listed roughly in order of how much the founder already knows about the person.

On matching platforms, Y Combinator runs Co-Founder Matching, which its page describes as free, with no equity taken. YC says the platform has made over 100,000 matches and lists active profile counts by city. A profile gets a founder into conversations. It does not replace the trial described below.

Former colleagues sit at the top of the table for a reason. The question that matters most, how this person behaves when things go wrong, is one the founder can already answer.

The first three conversations

Treat the early meetings as mutual diligence. The engineer is evaluating the founder at least as hard.

Conversation one: the problem. The founder explains the customer problem and the evidence. The engineer’s questions show how they think. Someone who jumps straight to the tech stack before asking who the customer is may be a strong engineer and a weak cofounder.

Conversation two: the work. Each person describes how they like to work: hours, location, how decisions get made, what happens when they disagree. Remote or in person, full time now or after a notice period, salary needs for the next 12 months.

Conversation three: the deal. Equity split, vesting, titles and who sits on the board. YC partner Michael Seibel’s essay on splitting equity among co-founders argues for equal splits because the work is ahead of the team. A founder who wants an unequal split should be able to explain why in a sentence an investor would accept.

Run a paid trial project

A trial project is the closest thing to working together without being married. Four to six weeks is long enough to see how the person handles ambiguity, scope cuts and a missed deadline.

A simple trial scope, written as a short agreement:

Trial project agreement (outline)

Parties:        [Founder A] and [Engineer B]
Term:           6 weeks starting [date]
Scope:          Build [one workflow] end to end for [named pilot customer]
Check-ins:      Twice weekly, 30 minutes
Compensation:   [cash amount / none]; no equity promised during the trial
IP:             All work product assigned to [Founder A, or the company
                once formed]; Engineer B confirms no employer claims it
Exit:           Either party may end the trial with written notice
After trial:    If both continue, incorporate and sign founder documents
                within 30 days

Two lines in that outline protect both people.

The IP line. Code written during the trial has to belong to the company, or to the founder for transfer to the company once it exists. An engineer with a current employer should check that employment agreement first. Employment agreements can include invention-assignment clauses, and work done on an employer’s laptop or time can create a claim. A lawyer should look at the clause before any code gets written.

The equity line. No equity is promised during the trial. Promising “10% if this works out” in a chat message creates an argument later about what was promised and when.

What to sign once both say yes

Once the trial works and both founders commit, the paperwork follows the same order as any new company:

• Incorporate, as a Delaware C corporation if the plan is to raise venture money.
• Issue founder stock to each founder under a stock purchase agreement with vesting.
• Have each founder sign an IP assignment covering prior work on the product, including the trial code.
• File each founder’s 83(b) election within 30 days of the stock purchase. The IRS instructions for Form 15620 set out the 30-day rule.

This is general information, not legal or tax advice.

When the search stalls

A three-month search with no candidate is a signal, not a failure. Three causes come up again and again:

• The evidence is thin. Engineers are pitched constantly. A founder with paying pilots gets a different response than one with an idea.
• The offer is an employee offer. Single-digit equity with a title of “CTO” reads as a hire.
• The search is too narrow. Only looking for a senior engineer from a famous company, when the product needs someone who ships quickly.

There are fallbacks. A founder can hire a contractor to build a prototype under a work-for-hire agreement with IP assignment, then use the prototype and customer results to recruit a cofounder. That costs cash and gives up nothing on the cap table, but it delays the moment when the company has a technical owner, and investors will ask who that owner is.

Red flags in a technical cofounder

None of these is disqualifying alone. Two or more together should end the conversation.

• Unwilling to sign an IP assignment or talk about vesting.
• Wants a large equity grant up front with no vesting.
• Cannot name a project they finished.
• Will only work nights and weekends with no plan to go full time.
• Dismisses the customer work as “sales stuff.”

The same list applies in reverse. A good engineer will be checking whether the non-technical founder brings something beyond the idea, and the honest answer has to be customers, distribution or money.