Pricing, the 48-hour trial, what happens when an engineer isn't a fit, payouts, and what we mean by โAI-native.โ Grouped by audience.
For founders
Booking, pricing, replacement.
How is this different from Toptal or Turing?
Two things: speed and structure.
Toptal/Turing are recruiters. They source candidates and you run a hiring loop, interviews, negotiations, contracts, monthly minimums. Average time to a working engineer: 3-14 days.
Cadence is a booking platform. You describe the role, the system auto-matches 4 vetted engineers, you take 30-min intro calls today, pick one or more, and they start work in 48 hours. Weekly billing, cancel any week. No interviews, no contracts, no notice periods.
How does the matching work?
When you submit a booking, our algorithm scores every available engineer on five dimensions: skill overlap (40%), rate compatibility (20%), timezone fit (15%), past founder ratings (15%), and current utilization (10%). Vertical match (e.g. fintech, healthtech) adds a bonus.
We then schedule 4 ร 30-min intro calls back-to-back at slots you picked, so you can decide on the same day.
What's the 48-hour free trial?
After intro calls you can pick one OR more engineers to start a 48-hour trial in parallel. They onboard, you observe how they work, ship something small. At the end of 48 hours you decide who to keep on weekly billing and who to release. You're not charged for any engineer you release.
Can I work with multiple engineers?
Yes. Pick more than one candidate after intro calls, each enters their own 48-hour trial in parallel. After the trial, keep any combination. Each engineer is on a separate weekly subscription so you can cancel them independently.
What happens if the engineer isn't a fit?
Two paths.
1. Mid-trial: release them with one click. No charge for trial work. We don't send another shortlist automatically, you can book a fresh request whenever.
2. Post-trial: hit "Replace at week-end" on the active engineer card. They wrap up at the end of the current week, and we shortlist 4 new engineers for intro calls tomorrow. The new engineer takes over Monday.
What does the engineer's day actually look like?
Async-first. Most engineers are in IST/EST/CET timezones; you set your preferred work hours during onboarding. They check in over Slack/Linear, attend a daily 15-min sync if you want one, and ship code. You give a ๐/๐ daily rating that takes 2 seconds. The data feeds the matching engine and engineer ratings.
Who are the engineers?
All engineers pass an AI-native voice interview (50% threshold) before they're matched to anyone. They self-report their stack, rate, and weekly availability. Top engineers (relevance score 90+) get priority on $1500โ$2000/week tiers. We verify workspace, internet (200+ Mbps), and time-zone before activation.
How does pricing work?
Four tiers, $500 (Junior, 0-2 yrs), $1,000 (Mid, 3-5 yrs), $1,500 (Senior, 5-8 yrs), $2,000 (Lead, 8+ yrs) per week. 40 hours dedicated by default; you can drop to 10 or 20 hours/week. Engineer keeps 80% of the weekly rate; we keep 20%. If you came via a partner link, partner keeps 10% and we keep 10%.
Can I cancel any time?
Yes. Hit cancel on any active booking. Service ends at the end of the current week (we bill the partial week). You can also pause indefinitely without ending the engagement.
Do I sign a contract?
No long-term contract. Standard SaaS-style terms of service apply. Engineers operate as independent contractors via Cadence, IP belongs to your company, NDA is enforced through our platform terms, and we handle the contractor agreement on the engineer side.
For engineers
Joining, matching, payouts.
How is this different from Upwork or Toptal?
Upwork: you bid for gigs, race to the bottom, freelancer-grade pay. Toptal: 4-week recruitment loop, 50%+ take-rate, you compete with 1000s for each placement.
Cadence: the system pushes pre-qualified opportunities to you. You confirm or decline. No bidding, no "sales calls." Take rate is 20% (you keep 80%). Weekly payouts every Friday.
What's the AI-native voice interview?
Three prompts, you answer in one voice recording (1โ3 min):
1. Walk me through a recent feature you built using AI tools.
2. How would you tackle a vague spec like 'build a Stripe-like dashboard'?
3. A time AI gave you the wrong answer and how you caught it.
Claude grades on AI-native fluency (35), communication (30), technical depth (25), and culture fit (10). 50/100 unlocks the platform. You can retake in 7 days if you don't pass.
How does matching work?
Founder posts a request โ our algorithm picks the top 4 engineers based on skill, rate, timezone, ratings, and current utilization. Those 4 get an opportunity offer (24h to respond). Accept it and you're scheduled into a 30-min intro call. Founder picks who they want for the 48-hour trial.
What's the relevance score?
Composite of: client ratings (40%), call-to-hire conversion (20%), on-time delivery (20%), response time to opportunities (10%), referral activity (10%). Stays at your interview score until you have real engagement data.
Relevance โฅ 90 = first-look access to YC/funded startup briefs and the $1500-$2000 rate tier. Below 70 = fewer opportunities.
How am I paid?
Weekly. Founders pay every Monday, you receive 80% of the gross every Friday (subject to a 1-week hold on first booking). Stripe Connect to your bank, Wise to 50+ countries, or PayPal. Minimum payout $50.
If you brought the founder via your referral link, you also earn 10% recurring of their weekly bookings, for as long as they stay on the platform.
What if a founder rates me negatively?
One negative rating: nothing happens, life is messy. Three negative ratings in a 5-day window: the founder gets a polite suggestion to request a replacement. They're under no obligation; many founders push through and the rating recovers.
If they do replace you, you wrap up at the end of the current week (paid in full) and the engagement closes. Your relevance score takes a small hit but recovers in 2-4 weeks of positive ratings on other engagements.
Can I work multiple engagements at once?
Yes, but the system tracks utilization. If you're on two full-time bookings (40 + 40), you'll get fewer opportunity pushes because the algorithm sees you as at-capacity. Quarter-time (10 hrs) or part-time (20 hrs) bookings stack better.
Do I have to accept every opportunity?
No. Decline as many as you want, we ask for a quick reason (rate, stack, capacity, etc.) so we can send better-fit opportunities next time. Declining doesn't hurt your relevance score; ghosting (no response in 24h) does.
What stack / vertical specialization is hot?
Currently most demand: Next.js + TypeScript fullstack, RAG/LLM apps (LangChain, Pinecone, OpenAI APIs), and Go services. Verticals: AI-native, fintech, SaaS B2B, health-tech.
If your stack is niche (Solidity, embedded, Unity), opportunities are lumpier, set a higher bar on what you accept.
What about IP, NDA, contracts?
All work-for-hire, IP belongs to the founder's company. NDA is enforced through Cadence platform terms (no separate signing). You operate as an independent contractor via Cadence. We provide a 1099 (US) or equivalent annual statement for taxes.