Recruiting from Scratch is a well-built, software-assisted contingency firm doing exactly what it claims: getting funded startups to a curated shortlist of permanent hires faster than the industry average. If you have approved headcount, a U.S.-based role, and you want a full-time W-2 engineer on your payroll, this is a serious option worth evaluating. The question isn't whether it works. The question is whether permanent placement is actually the right transaction for the stage you're in.

Recruiting from Scratch was founded by Will Sanders with a clear thesis: traditional recruiting firms win by volume, and volume is the wrong signal. Instead of flooding clients with 50 resumes, the firm built its own internal sourcing and matching platform to generate curated shortlists of three to five candidates per search. The firm focuses on technical and go-to-market roles at venture-backed startups and technology companies. That covers software engineering, hardware engineering, product design, product management, marketing, and accounting and finance. Every placement is a direct hire onto the client's own payroll, a permanent W-2 role, with no temp-to-perm or contract arrangements in the mix. The team operates fully remote across the United States, and they've placed engineers and GTM hires at companies including Scale AI, Coda, Eight Sleep, Newfront, and Bugcrowd, totaling over 300 completed placements by their own count. Sanders has framed the long-term vision publicly: building an intelligence layer over the recruiting process where every interaction becomes a learning loop to improve matching quality. That's not just marketing copy. It explains why the firm invested in proprietary tooling rather than running purely on LinkedIn Recruiter and spreadsheets like most contingency shops.

The headline metric is 29 days from intake to offer, compared to an industry average of 49 days. That's a 40% reduction in time-to-hire, which matters significantly at the Series A and B stage when a missing engineering hire is a shipping bottleneck, not just an org chart gap. The business model is straightforward. Recruiting from Scratch works on a contingency basis: clients pay nothing until a candidate is hired, and the fee is a percentage of the candidate's first-year salary. No retainers. No upfront costs. The firm also operates on Paraform, a recruiter marketplace where a partner case study highlights 36 placements, pointing to consistent output rather than one-off success. On the candidate side, the process includes a recruiter screen covering background, compensation, timeline, and role fit. Recruiters then stay involved through interview prep and offer negotiation, with no cost to candidates since client companies cover the fee.

Shortlist discipline. Most contingency firms send whatever cleared the ATS. Recruiting from Scratch sends three to five candidates who have been screened on compensation expectations, timeline, and role fit. That signals recruiter accountability. When you only get five names, you have real skin in the game on each one. Speed at the top of the market. Twenty-nine days average time to offer is verifiable, meaningful, and well below the industry norm. For a seed-stage team where every sprint month counts, the difference between a 29-day and a 49-day recruiting cycle is not trivial. It can be the difference between shipping a feature before a competitor does. Strong client portfolio. Scale AI, Coda, Eight Sleep, Newfront, Bugcrowd. These are recognizable, well-run companies with high hiring bars. Consistently placing into them isn't luck; it's repeatable pattern matching. Over 300 placements is a real number that reflects operational depth, not a firm still finding its footing. No upfront financial risk. The contingency model means a startup with tight burn can engage the firm without a retainer check on day one. You pay on success. That aligns incentives well and removes a common objection for early-stage founders watching every dollar.

The transaction ends at the signed offer. Recruiting from Scratch's job is complete the moment a candidate accepts. The full-time employment relationship, payroll, benefits administration, and compliance are yours from day one. That's the nature of any direct-hire model, and it's the right model for teams that want engineers fully integrated as employees. Just be clear-eyed that the vendor's accountability window is narrower than it appears. The fee structure rewards higher-comp placements. Because the fee is a percentage of first-year salary, the economics naturally align around senior, U.S.-based roles at market rate or above. This isn't a flaw; it's a feature for the client profile they serve. But if your headcount strategy involves flexible, non-headcount capacity or you're cost-sensitive on comp, that math deserves attention before you engage. Permanent placement is a one-way door. Once a candidate is hired onto your payroll, unwinding that relationship involves offboarding, potential severance, and internal process. That's true of all direct hiring, not unique to this firm. But if there's any ambiguity about whether you need a full-time headcount versus flexible engineering capacity for a defined scope of work, that question is worth resolving before you trigger a permanent placement search.

Recruiting from Scratch finds great permanent hires. That's real. But here's what happens when you close the search: you own the headcount, the payroll, the compliance, and the employment liability from the moment the offer is signed. The vendor's job is done. There's also a sourcing reality that no contingency firm will say out loud: the engineers you actually want for your most important work aren't applying anywhere. They're already working. Booked on something. Heads down, building. The same LinkedIn pools every agency queries surface the same candidates already in-market. Nextdev reaches the top 1% of AI engineers — the ones who aren't looking. We reach them directly, through our own outreach using LinkedIn response data, not keyword search over the same candidate pool everyone else is querying. We surface engineers who would never show up in a contingency search because they never entered one. Then we give each of them a real problem to build and watch how they work. Not a quiz. Not a take-home that gets passed to a friend. A real build, observed. That's how you know an engineer is genuinely AI-native: because you've seen them use it to think, not just to autocomplete. And when you want to engage one? We employ them. You get one contract and one invoice. Payroll, benefits, and compliance stay on our side. There's no new headcount req on your org chart, no offer negotiation, no employment liability to carry. If scope changes, you scale without a severance event. You're not picking from who signed up. You're getting the engineer who never would have.

Recruiting from Scratch is doing something genuinely well: bringing software discipline and shortlist precision to a contingency model that has historically rewarded volume over quality. The 29-day time-to-hire is real. The client logos are real. The no-retainer model reduces your entry risk meaningfully. If you have approved headcount for a permanent U.S.-based hire, a clear role definition, and you want a recruiter who will run point through offer close and send you five names instead of fifty, this firm deserves serious consideration. That's an honest recommendation. The fit breaks down when the engineering problem isn't a permanent headcount gap. If you need to move fast on a defined piece of work, want to test a hire before committing, need cross-border flexibility, or want to keep your headcount number flat while expanding output, direct-hire placement puts the burden of employment right back on your plate. That's not a failure of the firm. It's a mismatch between the transaction and the need. The most ambitious engineering organizations in 2026 aren't choosing between hiring more or hiring fewer. They're choosing between teams that multiply individual output through AI and teams that don't. The engineers who already know how to work that way are the ones already working. Finding them is a different kind of search.