How to find, evaluate, and close engineering hires at a startup

Executive overview

Hiring engineers is a startup's hardest recurring challenge — it's slow, repetitive, and demoralising, and most founders underinvest in it. Harj Taggar (TripleByte, ex-YC partner) covers sourcing strategy and closing; Ammon Bartram covers technical evaluation. Their data comes from interviewing 1,000+ engineers and observing candidates across multiple companies.

The core insight: interviews are far noisier than hiring managers believe — and structuring your process, not raising the bar, is what actually improves hires.

Building the sourcing funnel

• Treat hiring as a three-stage funnel: sourcing, screening, closing.
• Personal networks are the highest-quality source, especially for first hires — they de-risk both skill and working relationship.
• Run an exhaustive process: list every good engineer you know, make the ask regardless of availability, follow up rejections by asking for referrals.
• Scale this to your team via "sourcing parties" — 30–45 min, shared spreadsheet, everyone mines their own networks.
• Hiring marketplaces (TripleByte, Hired, Vetri) surface actively-looking candidates and compress time-to-hire; expect 15–20% fee per hire.
• LinkedIn/GitHub outreach requires personalised, evidence-based messages at low volume — high-volume spam has eroded response rates industry-wide.
• Use LinkedIn Recruiter Lite ($120/mo) + Connectifier to reach via email; email response rates significantly exceed in-platform messages.
• Job boards (Stack Overflow, AngelList, HN Who's Hiring) bring volume but require heavy resume filtering; stand out with a first-person founder story.
• In-person meetups are a long shot; hosting technical meetups at your office builds pipeline over time.

When and how to use recruiters

• Make at least your first engineering hire yourself before delegating to a recruiter — you need to feel the pain and develop the pitch.
• Start thinking about recruiters when you're hiring roughly one engineer per month and sourcing consumes more than 50% of your time.
• Start with a contract recruiter focused solely on sourcing (filling your calendar with calls); add a full-time in-house recruiter later to handle initial calls and on-site scheduling.
• Avoid agencies as a first step — 25–30% fees and misaligned incentives.

Reducing noise in technical interviews

• Inter-rater reliability across interviewers is just above 0.1 — roughly equivalent to random Netflix movie reviews. Interviews are noisier than hiring managers think.
• Trial periods are more accurate but only 20% of engineers will do them; the best engineers skew toward the 80% who won't.
• Decide which skills matter before designing the process: fast/iterative vs careful/rigorous; product-motivated vs tech-motivated; language-specific vs generalist; CS fundamentals vs product knowledge.
• Use structured interviews: same questions for every candidate, defined criteria per interviewer, centralised final decision-making. Free-form interviews feel better but are less predictive — there is no excuse not to use structure.
• Hide credentials (school, previous employer) from technical interviewers during coding evaluation; consider them only at the decision meeting. Interviewers demonstrably give halo-effect scores to credentialled candidates.
• Use good questions: multi-step problems with no single leap-of-insight, no specialised knowledge unless you've decided that matters, budget 3x your own solve time for the candidate.
• Ask four or more questions per session to average out noise.
• Calibrate on a candidate's peak performance, not their average — everyone looks stupid on the wrong question.
• A false negative (rejecting someone good) is as costly as a false positive, especially at early-stage companies struggling to source. Be less aggressive at pre-on-site screening if your pipeline is thin.
• For early-stage companies unsure what skills to hire for: prioritise productivity and ownership over code quality for first hires.
• Hiring outside your expertise: ask candidates to explain their answers in terms you can understand — genuine experts can do this.
• Reduce candidate stress: let them use their own laptop and tools, coach interviewers on soft skills, ensure every candidate — regardless of result — leaves wanting to work there.

Making offers and closing

• Speed is your biggest advantage over large companies. Big companies take weeks to deliver offers; be fast at every step from first contact to written offer.
• Train interviewers to be welcoming, not status-signalling. The top two candidate complaints: interviewers unfamiliar with their own questions; interviewers trying to look smarter than the candidate.
• Have a clear, specific answer about company culture — avoid the generic "open, transparent, collaborative" trifecta. Name trade-offs; it filters for candidates who are genuinely aligned.
• Involve team members and relevant investors after extending an offer; match the investor to the candidate's specific hesitations (e.g. someone leaving a big company should hear from someone who made that move successfully).
• Present full offer details upfront: salary, number of options, exercise price, vesting, cliff. Asking candidates to express interest before sharing compensation creates friction and reduces close rates.
• Check how well the candidate understands startup equity; provide follow-up resources if needed. Founders should understand option tax implications at the level candidates will ask about.

Competing against Google and Facebook

• Learning: startups give real decision-making authority with real consequences — unavailable at large companies for new hires.
• Career progression: the tech industry is the only sector where you can reach executive level by 35–40 by growing with a company.
• Opportunity cost: the current startup, team, and timing are unique; a safe big-company job is always available as a fallback.
• Mentorship pitch (junior hires): only make it if your team genuinely has senior engineers who will mentor — don't oversell.