CMC Markets Funded x Verichain
Hiring Intelligence

Matching, ranking and one record.

The CV and history of every candidate are captured once, then used at every stage after. Every application is parsed, matched and tiered on arrival.

Suggested for Senior platform engineerTop 40 of 3,159 reviewed
PR
Priya RamanKubernetes, Go, platform lead at a Series B · 7 yrs
Tier 1 · elite fit
JW
Jamal WrightDevOps to platform, Terraform, on-call ownership · 6 yrs
Tier 1 · elite fit
NO
Naomi OkonjoFull stack, strong infra, no Go in production · 5 yrs
Tier 2 · strong fit
DM
Declan MulhernData platform, Python and Spark, adjacent stack · 8 yrs
Tier 3 · borderline
Clients see the tier. The raw score is never shown.

Four places the model does work. None of them decide.

01

CV parsing

On every upload, education, experience, projects and skills are extracted into structured fields. The candidate reviews them before anyone else sees them.

02

Match scoring

Every new application is scored against the role and placed in a tier with written reasons. Rescoring is one click when the role changes.

03

Candidate suggestions

The whole pool is pre-filtered in the database, then the top forty are ranked by the model. Suggestions for a role load in well under a second.

04

Interview insights

Teams transcripts are summarised for the hiring manager and company owners only. No hire recommendation is produced, by design.

Fit is shown as a tier

Required skills, title overlap and experience band are scored in the database first. The top candidates go to the model with the role, and it returns a tier with reasons. Sorting and thresholds use the underlying number; people see the tier.

1 Elite fit
2 Strong fit
3 Borderline
4 Weak fit

How we use AI

Settled decisions, written down so you can hold us to them.

Advisory only Nothing auto-rejects, auto-progresses or auto-hires a candidate. Every stage change is made by a person.
No hire recommendation Interview summaries describe what was said. They do not say whether to hire.
Human review on request Any candidate can ask for a person to review an AI-assisted assessment. The policy is in our privacy terms.
One model, no fine-tuning Prompts only, one model for every call, structured output. Predictable behaviour and a simple cost shape.
Nothing leaves the path Candidate data goes to the model provider and nowhere else in the pipeline. No third-party enrichers feed it.

See a shortlist built from your own pool.

Bring one open role. We will run the matching live on the call.