Track record
We check delivery history with named referees: what shipped, what their managers would say about their code in review, and whether they would hire them again.
The Navigaite assessment
How we assessAI-native developers.
Anyone can put “AI-native” on a CV. We treat it as a claim to be tested: in a live session, on unfamiliar code, with a senior engineer watching. This is what a Navigaite shortlist means.
The problem with taking it on trust
Usage is not the same as judgement.
There is a wide gap between a developer who has used an AI tool and one who has rebuilt how they work around it. The first pastes output and hopes. The second specifies precisely, verifies everything, and catches the mistakes the model makes confidently.
A CV cannot tell you which one you are hiring. An interview question barely can. So we watch them work.
The four stages
A shortlist earned through evidence.
Technical interview
A senior engineer goes deep on the stack you need: architecture decisions, trade-offs, and debugging under pressure. AI does not remove the need for technical depth; it raises it.
Live working session
The candidate takes a real ticket on a codebase they have never seen, using their own AI toolchain. We watch how they build context, break the problem down, verify output, reject bad suggestions, and explain every line they accept.
The judgement conversation
We ask about the AI output they binned, and why. Developers who cannot name a time the model was confidently wrong have not been paying attention.
What gets a candidate rejected
The signals that stop a shortlist.
- Accepting generated code they cannot explain.
- No verification habit: no tests, review passes, or checks against the specification.
- One tool adopted once, with no habit of keeping up as the toolchain moves.
- Careless handling of secrets, credentials, or client code around AI tools.
- Weak communication. A contractor who cannot explain their reasoning cannot raise a team's bar.
After placement
Assessment does not stop at the shortlist.
We run structured reviews at week one, month one, and quarterly with you and with the developer. If something is off, we act before you have to raise it.
If a placement is not right, we replace it under the terms of the engagement. You should never have to chase us.
Want to see it applied to your brief?
Tell us the gap. We will keep the conversation short, useful, and specific to the team you are building.
Tell us the gap