Hire AI engineers on proof, not promises.
Patunay AI tests engineers on real client work and checks every result against a published rubric, in minutes. You meet three to five people with the evidence to show they can do the job.
- engineers assessed
- 369
- engineers assessed
- earn Verified
- 1 in 20
- earn Verified
- to a shortlist
- 10 days
- to a shortlist
- AI/LLM development3.4
- APIs & integrations3.1
- Client communication3.3
- Security & data2.4
Design partners hiring through Patunay
- AcmeAI
- Lumen Health AI
- Northwind AI
- Kalye Freight
- Bayanihan Labs
A resume tells you what someone claims. We show you what they did.
AI makes every resume look strong. So we stopped reading them as proof.
What the resume says
Claimed- “Built an agentic claims pipeline at scale”
- “Expert in REST APIs, Salesforce, and Guidewire”
- “Excellent communicator, client-facing”
No artifacts. No way to tell what they did versus what a tool wrote.
What Patunay shows
Patunay Verified- Retries with capped backoff and jitterclient.ts, lines 48–59, explained live at 03:10
- Spotted the ambiguous requirementAsked the client, proposed a default in the README
- Calm incident communicationAI interview transcript, cited in the score
Every score cites its evidence. Our team audits AI decisions and handles every appeal.
From job description to proven hire
A software-assisted service. We do the scoping, verification, and curation with you.
- 1Day 1
Tell us the role
Paste a job description. We turn it into a Role Signal: the competencies and weights that define success, which you approve.
- 2Days 2–9
We match on evidence
We score our verified engineers against your Role Signal, check hard constraints like time zone, and confirm each person is interested.
- 3Day 10
Meet 3 to 5 people
A shortlist with the reasoning behind every pick: code, a recorded walkthrough, and how they handled a client conversation.
- 4Days 30, 60, 90
We track the outcome
Short reviews tie performance back to the original evidence, so every match makes the next one better.
Five ways we watch someone work
Each step produces evidence for named competencies. Patunay AI scores every step against a published rubric and cites the evidence behind each score.
Built for the AI era
Candidates may use AI and must declare it. When integrity signals don't add up, Patunay AI asks the candidate a follow-up question instead of guessing. Any result can be appealed to a person.
- 45 min
Technical screen
Questions, then hands-on debugging: we hand over broken code, logs, and docs, and they fix it until the tests pass.
- ~6 hrs
Practical build
A realistic client scenario with a deliberate ambiguity. Commit history included.
const wait = Math.min( 2 ** attempt * 500, BACKOFF_CAP_MS, ); await sleep(wait + jitter());
- 10 min
Walkthrough
They explain their own code, then make a live change we choose.
- 20 min
AI interview
An AI plays a stressed client during a production incident.
- 20 min
Client update
A written update on a delay, for a non-technical reader.
“We stopped running take-home tests. Patunay's shortlist already had the code, the walkthrough, and the client conversation. Our first hire shipped a production workflow in week three.”
- 3–5
- people per shortlist, never a long list
- Minutes
- from last submission to a verified result
- 4.1 / 5
- average day-30 rating from companies
- 30/60/90
- day reviews on every engagement
Pay when someone starts
Role scoping, assessment, and shortlists are included. Choose how you want to work with the people you pick.
Placement
Hire the engineer onto your team.
- Role Signal and shortlist
- A replacement if your pick withdraws
- 30/60/90 reviews
Contract
Most chosenStart fast with an engineer on contract.
- Everything in Placement
- Flexible monthly terms
- Option to convert to a hire
Managed team
A small verified team with a Patunay lead.
- Everything in Contract
- Delivery lead and weekly reporting
- Scale up or down monthly
Engineers: let your work speak
Five self-paced steps over two to three weeks. Free. Everyone who finishes gets feedback and a development path, and verified engineers get a profile companies trust, shared only with permission.
Join with an invite codeUniversities: show outcomes
Onboard a whole cohort with one code. Track completion, average scores, and placements in aggregate. Students own their profiles.
Questions
Something else? Email hello@patunay.ai
Yes, and they declare it. We assume AI use and assess judgment: the walkthrough, the live change, and the interview show whether someone understands and owns their work.
Your next AI engineer, with the proof attached.
Send us the role. We'll send you three to five people worth meeting.