A startup hiring case study of 4 founders who hired 9 engineers
In this startup hiring case study, a 4-person seed-stage SaaS company in San Francisco, with no recruiter, HR or ATS, used The Cognitive's live AI video interviews to hire 9 engineers in 6 weeks. Customer-reported: $2,400 in total recruiting spend against $90K in agency quotes, and founder time on hiring down from 30 to 5 hours a week.
What did the seed startup report before and after?
The first 3 rows are the customer's own figures. The last 3 are our arithmetic on them, worked through line by line further down.
| Metric | Before | After | Change |
|---|---|---|---|
| Engineers hired (customer-reported) | 0 | 9 in 6 weeks | Inside the 8 to 10 plan |
| Total recruiting spend (customer-reported) | Agency quotes of $90K | $2,400 | $87,600 not spent |
| Founder hours on hiring per week (customer-reported) | 30 | 5 | 25 hours back each week |
| Spend per engineer hired (our arithmetic) | $10,000 at the quoted total | About $267 | Under 3% of the quote |
| Founder hours returned over 6 weeks (our arithmetic) | 0 | 150 | 25 x 6 weeks |
| Company headcount (our arithmetic) | 4 | 13 | More than 3 times the size |
Figures reported by the customer in our case study record dated 2 April 2026.
Who is the customer, and what did they have to work with?
The customer is a seed-stage SaaS startup in San Francisco, named in our record only as Seed-Stage SaaS Startup. When it came to us it had 4 people and had just closed a funding round. The plan was to hire 8 to 10 engineers quickly, enough to more than triple the company, with an MVP waiting on those engineers.
It had no recruiter, no HR person and no ATS. Every screening call fell to the founders, who were spending more than 30 hours a week on them. That is most of a founder's working week going on first conversations, at the stage when product work and investor updates also need that same founder.
We had no recruiter, no process, nothing. We set this up on a Monday, posted jobs Tuesday, and had 3 offers out by Friday. Co-founder and CTO, Seed-Stage SaaS Startup, San Francisco
Why didn't the founders hire a recruiting agency?
They asked. Recruiting agencies quoted around $90K to fill the roles. Spread across the 9 engineers the company went on to hire, that is $10,000 a seat. For a company that had just raised a seed round, it meant spending a large share of new cash before a single engineer had written a line of code.
Doing it themselves had a cost too, paid in hours. At 30 hours a week of screening calls, the founders had little time left for the product those engineers were being hired to build. The CTO put the stakes plainly: "At our stage that speed is the difference between shipping and dying."
The market made the hours worse. Gem's 2026 benchmarks show smaller companies advance 25% of applicants to pre-onsite stages but hire only 0.3% of them, roughly 1 hire per 333 applicants. For a founder doing every first conversation personally, that early funnel is where the 30 hours a week went.
How does 9 engineers in 6 weeks compare with typical startup hiring?
Published 2026 benchmarks, read 5 October 2026, next to the startup's reported figures. Benchmarks cover all company sizes unless stated.
| Measure | Typical | This startup | What it means |
|---|---|---|---|
| Application to accepted offer, technical roles | About 48 days (Ashby 2026) | 3 offers out in the first week; 9 hires in 42 days | The whole team hired inside the time one typical offer takes |
| Interview time per engineering hire | 24.7 hours, all interviewers (Ashby 2026) | 5 founder hours a week, about 3.3 per hire | About 222 hours at the benchmark for 9 hires, against 30 |
| Small-company funnel | 25% reach pre-onsite, 0.3% hired (Gem 2026) | Founders met only candidates above 80 out of 100 | The wide early funnel moved off founder calendars |
| Cost of the 9 hires | Agency quotes of $90K | $2,400 | About $267 per engineer |
Per-hire hours and the 222-hour figure are our arithmetic: 5 x 6 / 9, and 9 x 24.7.
How long does it take a startup to set up AI interviews?
Setting up a role: answer a few questions about the job, and the role is ready to invite candidates to.
The founders set everything up in one afternoon, with no recruiter, HR team or ATS to configure first. The quote gives the pace: set up on a Monday, roles posted on Tuesday, and 3 offers out by Friday. Signup is self-serve, so there was no sales call to wait for.
How did 4 founders run hiring without a recruiter?
The setup dates from April 2026 and the product has changed since. These steps describe how the same approach runs in The Cognitive today.
- Write the bar for each role Give each role a few weighted criteria and say what a weak, middling and strong answer looks like. The founders tightened their technical questions after the first reports came back too generic. Criteria we suggest for a first engineer: ownership 30, technical depth 30, tradeoff judgment 20, debugging 10, communication 10.
- Post the roles where engineers look The founders posted on LinkedIn and job boards themselves. The Cognitive is not an ATS and does not post jobs, so the listings stay wherever you already advertise.
- Bring applicants in and bulk invite them With no ATS, applicants go in by CSV upload. Select the batch and invite everyone in one go; candidates book their own slot without creating an account and get email reminders. Upload and invite each day's applicants that evening, so a Tuesday posting can turn into Friday offers, as it did here.
- Let the AI hold the first conversation The AI interviews each candidate live over video, asks technical and behavioral questions and follows up on thin answers. Candidates explain their reasoning out loud; there is no code editor. Set engineering roles to 20 minutes, the longer of the 2 lengths, so the AI has room to follow up on a project story.
- Founders read reports and meet only the top scorers Each report gives 1 to 5 per criterion, a weighted score out of 100, overall written feedback, a transcript and a recording. The founders met only candidates above 80 out of 100. Split the reading: the CTO reads technical depth and watches recordings, the CEO reads ownership scores and overall feedback.
- Founders run the final conversation and close The remaining 5 hours a week went on final conversations, offers and closing. The AI suggests a verdict and nothing is rejected automatically, so every decision stayed with the founders. Make the offer within days. 3 of the 9 hires said speed was why they chose this startup.
- Search when a role draws too few applicants For a hard seat, describe the engineer in plain English, search ~900M public profiles, reveal verified emails and phone numbers, and invite the people you shortlist to the same interview. A brief like: founding backend engineer, TypeScript and Postgres, early-stage startup experience, San Francisco Bay Area.
How did total recruiting spend come to $2,400 instead of $90K?
The record describes the $2,400 as subscription cost rather than placement fees, and that difference in pricing model explains most of the gap. An agency charges per hire, so its bill grows with every engineer placed. An AI Interview plan includes a monthly interview allowance, with extra interviews billed at the plan's overage rate. The cost follows how many interviews you run rather than how many people you hire.
Divided across 9 hires, $2,400 is about $267 per engineer, against $10,000 per engineer at the agency quote. The company kept $87,600 of the quoted amount, and $2,400 is under 3% of $90,000.
Founder time is the other half of the bill. Price the 150 founder hours returned at the BLS 2025 median for software developers, QA analysts and testers, $64.44 an hour (BLS, read 5 October 2026), and they are worth about $9,666, four times the $2,400 spent. Founder hours at a seed startup arguably cost more than that, since they also run fundraising and sales.
The working, line by line
Inputs are from the customer's record, plus BLS pay and Ashby's benchmark. Everything after an equals sign is our arithmetic.
SPEND PER ENGINEER HIRED Agency quote: $90,000 / 9 = $10,000 Actual: $2,400 / 9 = about $267 SPEND NOT USED $90,000 minus $2,400 = $87,600 $2,400 / $90,000 = 2.7%, under 3% of the quote FOUNDER HOURS ON HIRING Before 30 a week. After 5 a week. 30 minus 5 = 25 hours back each week. 25 / 30 = 83% less. 25 hours x 6 weeks = 150 founder hours. 150 x $64.44 (BLS median hourly pay) = about $9,666. Founder hours per hire after: 5 x 6 = 30 / 9 = about 3.3. HIRING PACE 9 engineers / 6 weeks = 1.5 a week 9 hires in 42 days, against Ashby's 48 days for one technical offer COMPANY SIZE 4 + 9 = 13 people 13 / 4 = 3.25 times the starting size AGAINST THE PLAN OF 8 TO 10 ENGINEERS 9 hired, inside the range.
What did the founders do with 150 hours back?
Founder time on hiring fell from 30 hours a week to 5, so 25 hours came back every week. Over the 6-week push that is 150 hours, close to a month of one founder's full working weeks. The record's lesson is that at seed stage the real cost of recruiting is founder attention rather than the fee.
The record also reports that the startup shipped its MVP 2 weeks ahead of schedule, with 9 more engineers and founders who were back on the product. Speed also helped the startup win people. 3 of the 9 engineers said how fast the process moved was why they picked this startup over larger companies.
A founder's week, before and after. Before: 30+ hours of screening calls, each one a first conversation the founders ran themselves, with no recruiter or ATS behind them. After: about 5 hours, spent reading reports from candidates above 80 out of 100, running final conversations and closing. The other 25 hours went back to the MVP that shipped 2 weeks early.
What did candidates experience in the AI interview?
A live, two-way AI video interview: the candidate checks camera and microphone, the AI asks and follows up, and the room closes on time.
Each candidate met the AI on video at a time they booked themselves. The rubric stayed fixed for the role, while the questions adapted to each answer, so a candidate who mentions a payments system, for example, gets asked how they handled failures in it. The invitation told candidates the interview is run by AI.
Which example questions suit a seed-stage engineering interview?
Examples you could add to a role as your own questions. These are illustrations, not this customer's questions. The AI asks added questions in its own order and adds follow-ups.
Tell me about something you built end to end with no one to hand the hard parts to.
Names the decisions they made alone, what they got wrong, and how they found out. Talks about users or customers, not only the code.
What would you build differently if you started it again today?
You have 2 weeks to ship a feature that would take 6 to do properly. Walk me through what you cut.
Separates what must work on day 1 from what can wait, says how they would tell users, and names the shortcut they would refuse to take.
Which of those shortcuts would you go back and fix first?
Production is slow after a deploy and there is no one else on call. What do you check first?
Starts by checking what changed and whether to roll back, reads the metrics before guessing, and explains how they would confirm the fix.
The rollback did not help. What next?
What does this case study not show?
- One startup, one 6-week hiring push, so use it to plan your own first hires.
- The interview has no code editor; add a coding exercise as a final round if a role needs one.
- The Cognitive does not post jobs or track a pipeline, so roles were posted and tracked by the founders.
- The setup dates from April 2026. The product has changed since, and this page describes today's workflow.
What still needed the founders?
Every hire. The founders read the reports, met the top scorers, made the offers and closed the candidates. The Cognitive does not send offers, schedule human interviews or track a pipeline, so a startup with no ATS still needs a simple place to keep track of who is at which stage, even if that is one shared sheet.
Calibration needed them too. The first reports were too generic for what the team cared about, so the founders rewrote the technical questions and made the candidate instructions clearer. Only after that did they trust the score enough to let it decide who they spent their 5 hours on.
Which startups does this approach fit?
Do
- Founders hiring several engineers at once with no recruiter on staff.
- Roles posted publicly that draw more applicants than founders can call.
- Teams that can write down what a strong answer looks like.
- A product date that slips if hiring takes months.
Do not
- 1 or 2 hires filled from the founders' own network.
- Roles where a hands-on coding exercise must come first; add one as a final round.
- Founders who want an agency to close candidates for them.
Try it on your first engineering role
Set up one role this afternoon, upload the applicants you already have and read the reports before you take another screening call. Start free at https://app.thecognitive.io/signup.
Where to go next
Hiring software for small business · Founder hiring playbook · Hire software engineers with AI · AI recruiting for tech hiring · Pricing · All case studies
Sources and notes
Figures and quote reported by the customer in The Cognitive case study record for Seed-Stage SaaS Startup, San Francisco; setup dated April 2026 (record dated 2 April 2026). Derived figures are our arithmetic and are labelled. The example questions are illustrations written for this page. The workflow is described as the product works today.
Benchmarks, all read 5 October 2026: Ashby 2026 Talent Trends, recruiter productivity (about 48 days application to offer for technical roles, 24.7 interview hours per engineering hire); Gem 2026 Recruiting Benchmarks (smaller companies: 25% of applicants to pre-onsite, 0.3% hired); BLS Occupational Outlook, software developers ($64.44 median hourly pay in 2025 for software developers, QA analysts and testers).
Frequently asked questions
Where can I find a startup hiring case study with real numbers?
This page is one: a 4-person seed-stage SaaS startup in San Francisco that hired 9 engineers in 6 weeks using The Cognitive's live AI interviews. Every figure is labelled as customer-reported or as our arithmetic. More customer stories are at thecognitive.io/case-studies.
How does a startup hire engineers without a recruiter?
Hand the first conversation to software and keep the decisions for the founders. This startup posted roles itself, uploaded applicants, bulk invited them to live AI interviews and met only candidates scoring above 80 out of 100. Founder time on hiring fell from 30 hours a week to 5.
What is the best hiring software for a small business?
It depends on where your time goes. If screening calls eat the week, an AI interviewer such as The Cognitive takes them over while you keep posting jobs where you already do. If you need job posting and pipeline tracking, you will also want an ATS, because The Cognitive is not one.
Are there affordable AI recruiting tools for startups?
Yes. The Cognitive is self-serve with published monthly prices: AI Interview starts at $99/month and AI Sourcing at $49/month, on monthly plans you can cancel anytime. The free trial includes 2 AI interviews and 100 sourcing credits.
How much do recruiting agencies charge startups?
This startup was quoted around $90K to fill its engineering roles, which works out to $10,000 for each of the 9 engineers it hired. Agency fees are usually charged per placement, so the bill rises with each hire. AI Interview plans include a monthly interview allowance instead, with extra interviews billed at the plan's overage rate, so the number of people you hire does not change the bill.
How long does it take to hire engineers at a seed-stage startup?
This team hired 9 engineers in 6 weeks, about 1.5 a week, and had its first 3 offers out in the first week. For comparison, Ashby's 2026 data puts technical roles at about 48 days from application to accepted offer (read 5 October 2026).
How many applicants does a startup need per engineering hire?
Gem's 2026 Recruiting Benchmarks show smaller companies hiring 0.3% of applicants, about 1 hire per 333, while advancing 25% of applicants to pre-onsite stages (read 5 October 2026). That wide early funnel is the part a live AI interview takes off founder calendars.
How much founder time does hiring take at a seed startup?
This startup's founders spent 30+ hours a week on screening calls before, and 5 after. Ashby's 2026 data shows 24.7 hours of interview time per engineering hire across all interviewers (read 5 October 2026). Here founders spent about 3.3 hours per hire by our arithmetic.
What should the first engineers at a startup be interviewed on?
Ownership and tradeoff judgment as much as technical depth, because nobody is there to hand the hard parts to. Our suggested rubric: ownership 30, technical depth 30, tradeoffs 20, debugging 10, communication 10. See our software engineer interview questions for more prompts.
Can AI interview software engineers?
Yes, as a spoken technical interview. The AI asks technical and behavioral questions live over video and follows up when an answer is thin. Candidates explain their thinking out loud, and there is no code editor, so add a coding exercise in a later round if you need one.
Does the AI decide who gets hired?
No. Each report suggests a verdict from the weighted score, and a person accepts or overrules it. Nothing is rejected automatically. In this case study the founders made every call and ran every final conversation.
Do I need an ATS to use The Cognitive?
No. This startup had none and uploaded applicants by CSV. If you do use one, The Cognitive connects to 60+ ATS integrations and writes a note with the score, summary and report link back to candidates imported from it.
Do candidates prefer a fast hiring process?
Many do. 3 of the 9 engineers this startup hired said the speed of the process was why they chose it over larger companies. The founders had 3 offers out within a week of setting the process up.
Explore: Recruitment agency fees · An affordable hiring system · How the AI interviewer works · Software engineer interview questions · SaaS case study: 18 engineers · What is cost per hire