Your first software engineering hire, your founding engineer, might be the most important decision you make this year, and you are making it with no recruiter, no talent team, and a budget that does not stretch to either. Contingency agencies in the US charge 15 to 25 percent of first-year base, with 20 percent the working benchmark and a premium on senior engineering, which we break down role by role in the honest math behind the 20 percent. On a $180,000 senior engineer, 20 to 25 percent is $36,000 to $45,000 for one hire. For an early-stage company, that is a month of runway spent on a finder's fee.
So you do it yourself, which is also what founders who hired well early tend to have done. Brian Chesky told a Y Combinator audience in October 2014 that he had interviewed Airbnb's first 300 employees himself, a figure he had raised to 400 by the time he revisited it in 2024, and his stated regret was stopping too soon rather than starting. The hard part is not whether to do it yourself. It is knowing where to look, what to offer, and how to reach the people who are not answering job posts.
This is the playbook for that, written for the founder doing the hiring, not for a recruiting team you do not have.
The three default options, and where each one runs out
Most founders reach for the same three channels, and each one has a ceiling.
Your network is the standard first move, and the standard advice is emphatic about it. Y Combinator's own guide to hiring your first engineer, written by Harj Taggar in 2018, ranks channels best to worst and puts personal networks first, telling founders it is "where you should exclusively focus your energy" and that "for your first three engineering hires I'd recommend focusing exclusively on personal network hiring."
The case for it is real. Ashby, analysing 54 million applications across 93,000 jobs, found referrals make up about 1 percent of applications but roughly 18 percent of hires, and that 52 percent of referred candidates clear initial screens against 35 percent overall. On retention, the best evidence is academic rather than vendor: Burks, Cowgill, Hoffman and Housman, publishing in the Quarterly Journal of Economics in 2015, found referred workers 10 to 30 percent less likely to quit across nine large firms. Read that as evidence about the mechanism rather than a startup benchmark, because nine large employers are not a five-person company.
Here is the part the advice does not prepare you for. Ashby's February 2026 look at 1,200-plus VC-backed startups found referrals produce only 15 percent of startup hires, below the 18 percent all-company average, while the smallest startups get about 30 percent of their hires from sourcing. Ashby flags the gap itself as surprising given how much early founding teams lean on their networks. Read those two things together and the conclusion is not that YC is wrong. It is that your network runs out faster than you expect, and the founders who keep hiring are the ones already doing outbound by hire three or four.
Job boards and startup marketplaces are where you go next. Wellfound (formerly AngelList Talent) lets you post for free and is full of people who specifically want to work at startups, understand equity, and expect the ambiguity. Y Combinator's Work at a Startup is similar for the YC orbit. The ceiling here is that everyone on these platforms is actively looking, which means two things: you compete with every other startup for the same active candidates, and you never see the engineer who is heads-down and happy at their current job, which is often exactly the person you want.
A contingency agency removes the work but reintroduces the cost you were trying to avoid. For a funded company filling a hard senior role on a deadline, it can be worth it. For your first few hires on a tight seed budget, the math rarely works. Worth knowing what you are comparing against, though: SHRM's 2025 benchmarking puts the average internal cost per non-executive hire at $5,475, so doing it yourself is cheaper than an agency but it is not free, and the difference is mostly your hours. If you are weighing a fractional recruiter instead, we compared the options in in-house versus freelance recruiter for a startup.
| Channel | Cost | What you get | Where it runs out |
|---|---|---|---|
| Your network | Free | High-trust, high-retention hires | Runs dry after the first hire or two |
| Startup marketplaces (Wellfound, YC) | Free to post, paid tiers optional | Active, startup-minded candidates | Only active people; you compete for the same pool |
| Contingency agency | 15-25% of first-year base, 20% typical | Someone else does the work | $36K-$45K per senior hire; rarely fits a seed budget |
| Direct outreach to passive engineers | Your time, plus a sourcing tool | People who aren't on any job board | Hard to find them and know who's reachable |
The gap nobody hands you: the engineers who aren't looking
Stack those three options together and you have covered everyone who is actively in the market. That is the trap. The strongest engineers are usually not in the market. They are employed, busy, and not refreshing a job board. They will never see your Wellfound post, and they are not in your network yet.
Reaching them is what a recruiter would normally do, and it is the one part of the job that does not transfer cleanly to a founder, because it has two hard sub-problems. First, finding them at all, since they are scattered across GitHub, past projects, and companies you would not think to look at. Second, knowing which of them is even worth approaching, because emailing 200 happily-employed engineers cold is a great way to burn a week and your reputation.
GitHub is the closest free version of this. Active contributors to projects in your domain are real engineers with public, readable work, and a thoughtful note about their actual code lands far better than a templated pitch. It works, but it is slow, and it tells you nothing about whether the person is open to moving or what they would expect to be paid. If you want the technique in more depth, passive candidate sourcing covers it properly, including why the widely quoted "70 percent of the workforce is passive" framing misleads people.
Know what to offer before you make the offer
The fastest way to lose a candidate you spent weeks sourcing is to open with a number that is wrong. Founders without a comp background routinely misprice the first few offers, either lowballing a senior engineer into walking away or overpaying out of fear and breaking the band for everyone hired after.
You need a real read on the market for the specific role, seniority, and location before the conversation gets to numbers. Public salary data is noisy and self-reported. Live market benchmarks, like the ones in Glozo's market intelligence, are built from active postings and give you a defensible range to anchor on. Your stack matters here more than founders expect, because supply and pay diverge sharply between languages and frameworks: we mapped that in hiring cost by tech stack. Walk into the offer conversation already knowing the band, and have it pre-approved with your cofounders, because in a competitive market the founder who can say yes the same day wins the candidate who is also interviewing somewhere slower.
Your real advantage is variance, and it cuts both ways
Founders are usually told that speed is their edge over big companies. The data says something more uncomfortable and more useful.
Ashby's 2026 benchmarking, drawn from 54 million applications, found that smaller organisations show a wide distribution on time to fill rather than a shorter one. Their words: startups "can move very quickly, outperforming larger organizations, but are also at risk of having hiring efforts drag out for much longer," and that variability only narrows as a company grows, with timelines starting to stabilise around 100 employees. Senior roles take about 37 percent longer than junior ones, and technical roles run roughly 15 days longer than business roles.
So you are not reliably faster. You are less predictable, and which side of that distribution you land on is mostly determined by whether you have a pipeline before you need one. The startups that beat a big company to a candidate do it by compressing the decision, not the search: interviews led by the founder, a comp band already agreed with cofounders, an offer the same day. The ones whose searches drag for five months are usually the ones that started sourcing the week the role opened.
Practically, that means two habits. Keep a running list of engineers you would hire, added to whenever you meet one, long before there is a req. And when someone strong is in your process, remove every scheduling gap you control. A strong engineer choosing between you and a safer name is often swayed by the founder who clearly cares and clearly moves.
Where Glozo fits for founders
Glozo is the self-serve version of the part you cannot do alone: reaching the engineers who are not on a job board, and knowing who is worth your time before you spend it.
It works on three signals a job board does not have. It reads candidate data from 30-plus sources and more than 10 million market signals a month, so you can find engineers by what they have actually built rather than who happened to post. It gives you a Market Value estimate per person, so you know the comp band before you reach out. And its Open to Offers signal points to people who are likely receptive, even if they never set themselves to "open to work," so your outreach goes to the few who might say yes instead of the many who never will. No percentage-of-salary placement fee, no marketplace bidding war for the same active candidates. You run it yourself, the way you already run everything else this early.