Know who can do it, not who says they can
Talynce rates every skill on the evidence in the resume, scores every interview answer on the same rubric, and reads the code candidates write. Each number comes with the reasoning behind it.
Free for 1 user · No credit card required
Neha Kulkarni
Sr. Data Engineer · 4 answers scored
Overall is the plain average of the scored answers.
- Q1 · Data pipeline you built86
- Q2 · Ensuring data quality78
- Q3 · Coding: optimise the script82
- Q4 · Explaining it to a client64
- rating on every resume skill
- 0–10rating on every resume skill
- score on every answer
- 0–100score on every answer
- bands in the scoring rubric
- 9bands in the scoring rubric
- certificate levels for candidates
- 5certificate levels for candidates
Why it matters
Resumes claim. Interviewers differ.
Two things make a shortlist hard to defend: every resume lists the same skills, and no two people judge an answer the same way.
Today
- Every resume lists the same skills, with nothing to tell them apart
- Two interviewers, two standards, two sets of notes
- Coding ability taken on trust until the client's own round
- No way to tell whether an answer was the candidate's own
With Talynce
- Skills rated on where and how they were used
- One rubric, and a written justification for every score
- Code written in the browser and scored on what it does
- Integrity signals alongside every result
Skill ratings
A skill in a header is not a skill led on a project
Each skill Talynce finds in a resume is rated from 0 to 10 on the evidence for it, and shown as a bar on the candidate's Skills tab.
- Where it appears: a skills list, or real project work
- Ownership: words like built, led and designed
- How often it recurs across roles, and how recently
- Whether it is tied to a measurable outcome
- Java9/10
Led two builds, used in every role, current
- Spring Boot8/10
Owned services across three roles
- Microservices7/10
Designed one, recent, with an outcome
- AWS5/10
Used in one project
- Kubernetes2/10
Appears in the skills list only
Answer scoring
Every answer scored on the same nine-band rubric
Spoken answers are transcribed, then scored from 0 to 100. The overall score is the plain average of the answers that were scored.
- A justification of two to four sentences for every score
- Up to three strengths and three concerns per answer
- The same questions and the same rubric for every candidate
- Answers are scored in English
Describe a data pipeline you built and the problems you solved.
“We were loading about forty million rows a night and the job kept timing out. I split it into partitioned loads and added checkpoints, so a failure restarts from the last partition…”
Justification
A specific problem, the candidate's own decision and a measurable result. Depth on failure handling is strong; monitoring is mentioned only briefly.
Strength: owned the redesign end to end
Concern: little said about monitoring
Coding questions
Watch them write the code
Turn any question on a job into a coding question. The candidate writes the answer in an editor in the browser, inside the same interview.
- Set the prompt, the language, starter code and instructions
- Scored 0 to 100 on correctness, structure, edge cases and fit to the prompt
- A written justification with every score
- Large pastes into the editor are logged for the reviewer
Integrity signals
A score you have a reason to trust
Each interview carries a Trust Score from 0 to 100 and a risk band of low, medium or high, built from what happened during it.
- Tab switches, time away and fullscreen exits
- A missing face or more than one, in browsers that can detect faces
- A reference selfie taken at the start, for the reviewer
- Always shown as decision support, never as proof of misconduct
Trust Score 88
Starts at 100. Each signal takes a fixed amount off.
- Tab or window switches2
- Time away from the interview14 s
- Fullscreen exits0
- More than one faceNone
- Large paste in a coding answerNone
Decision support only, not proof of misconduct
Certificates
Candidates can prove a level before you ask
In the candidate portal, candidates take skill-level interviews built from their own resume and earn a certificate they can share.
- Five levels, from Beginner to Expert
- Fresh questions on every attempt, and 80% to pass by default
- The certificate shows the level, score, date and primary skills
- Shared by link, with the interview recording behind it
- 4
- 5
Certificate of Competence
Rohit Bansal
Intermediate · scored 84 · pass mark 80
Also included
Questions that fit the role
Drafted from the job description
Talynce writes five spoken questions from the job's title, skills and description. Edit, reorder or replace them.
Or drafted from the resume
Create a resume-based interview for one candidate, with no job attached, and review the questions before the invite goes out.
Re-analyze at any time
Re-analyze ratings re-scores a candidate's skills against their stored resume and replaces the old ratings.
How it works
From a question to a score you can defend
- 1
Set the questions
Generate them from the job description, edit them, and switch on Coding for the ones that need it.
- 2
The candidate answers
They take the interview from a link, speaking their answers and writing code in the browser.
- 3
Each answer is scored
Every answer gets a score, a justification, strengths and concerns. Integrity signals are recorded alongside.
- 4
You review the evidence
Read the scores beside the recording and the transcript, then decide who moves forward.
Your team stays in charge
Numbers with their reasoning attached
An assessment is only useful if a recruiter can check it and a client can believe it.
No hire or no-hire verdict
Talynce reports scores and the reasons for them. It does not recommend who to hire. That call is yours.
Careful about integrity
Signals are reported as signals. The product never says a candidate cheated, and a low Trust Score is a prompt to look closer.
One standard for everyone
The same questions and the same rubric, whoever the candidate is and whichever recruiter sent the link.
The rest of the pipeline
Assessments is one stage of four
Each one hands its results to the next, against the same candidate and the same job.
Each skill is rated from 0 to 10 on the evidence in the resume: where it appears, whether the candidate owned the work, how often it recurs, how recent it is and whether it is tied to an outcome. A skill that appears only in a skills list rates 1 or 2.
Score the next interview on evidence
Free for 1 user, with 10 AI video interviews in every billing period.