Register

For technical recruiters

Hire on evidence, not keywords

TalentGraph reads what each candidate has actually done, checks it against the public record, and writes the interview questions worth asking. No technical tests, no AI interviews.

Free for 14 days · No credit card · The trial ends automatically

Hosted in the EU · GDPR · Encrypted connections

See a sample candidate review ↓
Fictional demonstration

Lena Hartmann

Senior Backend Engineer (Go) · Berlin, Germany

Profile fit Strong

Our judgment

Can do this job: yes · at the job's level

Builds and runs Go routing services for a freight network, with Kafka at the centre of the work.

  • Runs Go services in production verified
  • Kafka consumers with replay verified
  • Kubernetes operations ask on the call
  • No technical tests
  • No AI interviews
  • No deep technical knowledge needed
  • Evidence, not keyword matching

How it works

TalentGraph judges the work, not the word.

01

Read the work

It reads what each candidate has actually done, not just the keywords on the CV.

02

Check the record

It checks claims against the public record where one exists, and shows you the evidence.

03

Ask better questions

Interview questions about operations instead of vendor buzzwords, with what a strong and a weak answer sounds like.

Start a free trial → See a sample candidate review

See it in action

From job description to a search that finds the hire

Start from your workspace

Press Post a role. Each role keeps its shortlist, your uploaded candidates and the assessment document together.

TalentGraph workspace with no roles yet and a Post a role button

Tell us about the role

Add the role title, location and work mode, then paste the job description or upload it.

New-role form with the role title Senior Web Developer, the location Bristol, United Kingdom and the Hybrid work mode filled in

Paste or upload the JD

Fill in the details and paste the full job description, or upload it as a PDF, DOCX or TXT file.

New-role form filled in for a Senior Web Developer, hybrid, with the job description pasted and the upload field below it

Add what the JD leaves out

Business context and success outcomes, what matters most in the search, and the companies we must not approach.

Business context, what matters most and company-that-is-hiring fields, with the Continue — read my JD button

TalentGraph reads your JD

It separates the must-haves from the nice-to-haves and checks what the JD leaves out. Usually under a minute.

Progress screen: Reading your job description, step 1 of 3

See what would hold the search back

A search-readiness score, a warning before you post, and the JD rewritten the way we would post it.

Help with the JD screen showing 88 percent search readiness, a before-you-post warning and a rewritten JD

Decide what gates the search

Must-haves gate the search. Nice-to-haves only refine the ranking and never exclude anyone.

Skills split into must-haves, with React as the leading skill, nice-to-haves, and the job titles the search looks for

See what you get

What a candidate review looks like

Fictional demonstration: the role, the people, their employers and every page named are made up. The cards show the same parts as a real review.

Fictional demonstration

Lena Hartmann

Senior Backend Engineer (Go) at Lindenfield Freight · Berlin, Germany

Profile fit Strong

Our judgment

Can do this job: yes · at the job's level

What the public record shows

Builds Go services for freight routing route-planner repository: Go, commits from 2023 to 2026 verified
Kafka consumers with replay event-replay README describes idempotent consumers and the replay tool verified
Kubernetes operations No public record found under this name not verified

What to check on the call

Kubernetes operations: only their own words say it. Ask for one concrete piece of work.

Fictional demonstration

Marta Soler

Backend Engineer (Go) at Quayside Parcel Co. · Valencia, Spain

Profile fit Promising

Our judgment

Can do this job: probably · at the job's level

The profile describes Go and Kafka work at the job's level, but nothing outside the profile confirms it yet.

What the public record shows

Go services in production No public record found under this name not verified
Event streaming (Kafka) No public record found under this name not verified

No independent record found; score capped at 50.

What to check on the call

Go services in production: only their own words say it. Ask for one concrete piece of work.

Open the full-length sample reviews →

The problem

Matching words doesn't mean matching talent.

Job descriptions have become wish lists. AI helps candidates write CVs that match.

Too many technologies to know

Technical recruiters are now expected to know several technologies, several clouds and every vendor's name for the same thing, and to judge exact answers in a field that is not their own.

One misheard word, one lost hire

A call transcript that mishears AKS for EKS, run through an AI screener, turns an experienced engineer into a rejection. Both are Kubernetes.

The solution for recruiters

Tell experience from vocabulary, without deep technical knowledge.

TalentGraph reads what each candidate has actually done, checks it against the public record, writes interview questions around operations instead of vendor buzzwords, and shows the evidence.

By closing the gap between recruiters and hiring managers, it supports more meaningful hiring conversations, without adding technical tests or AI interviews.

Read the full story →
Two colleagues work through a problem at a glass board

Hiring managers need people who can solve problems, handle incidents and learn, not simply tick every technology box.

FAQ

Questions recruiters ask

What is TalentGraph?

TalentGraph is an AI-powered technical hiring assistant for recruiters. It reads what each candidate has actually done, checks it against the public record where one exists, explains the technical terms in plain language and writes the interview questions worth asking, with what a strong and a weak answer sounds like.

Who is TalentGraph for?

Recruiters who hire for technical roles, including those without deep technical knowledge, and the hiring managers they work with. It closes the gap between the two, so hiring conversations are about the candidate's actual work.

How does TalentGraph check a candidate's claims?

It compares what the profile says with what the public record shows, such as public code repositories or talks listed under the candidate's name. Each claim is marked verified, claimed, not verified or not shown, and the sources are listed. When nothing outside the profile confirms the claims, the score is capped.

Does TalentGraph use technical tests or AI interviews?

No. TalentGraph adds no technical tests or AI interviews, and it does not score call transcripts or keyword lists. A transcript that mishears AKS for EKS, which are both Kubernetes, can turn an experienced engineer into a rejection. TalentGraph judges candidates on evidence of their work instead.

Where is candidate data processed?

TalentGraph is hosted in the EU (OVHcloud, France), works within the GDPR and uses encrypted connections. Some AI and search providers may process data outside the EU under standard contractual clauses — see the Privacy Policy and the Terms of Service.

What does the free trial include?

The free trial lasts 14 days and includes 1 role, 10 sourced candidates and 20 CV checks. You do not need a credit card. The trial ends automatically: it never turns into a paid subscription and nothing is charged. You sign up with your name, your company and your email; there is no password.

See the evidence behind your next hire

Start a free trial and judge your next candidate on the work they have done, not the words on their CV.

Free for 14 days · No credit card · The trial ends automatically

Hosted in the EU · GDPR · Encrypted connections