Skip to main content

For assessors

Measure real capability — independently and fairly

Assessors are hoctienganh.vn's independent measurement layer: you rate speaking against calibrated rubrics without knowing who is speaking. Your ratings decide whether a capability is verified.

Attempt #A-2817Blind
Task
Handle a delivery complaint by phone
Target listener
Singapore partner, moderately formal
Rubric
Intelligibility · Task completion · Responsiveness
Hidden
Name, email, organisation, teacher, AI score

Shortcuts: 1–5 score · Space play/pause · ←/→ seek 3 s · Enter submit

Illustration of the rating queue.

At a glance

For
Teachers, language specialists, native or proficient listeners
Work
Blind rating of speaking against rubrics, remote
Quality
Calibration before live work and periodically after
Independence
Never rate your own learners or items you wrote
Schedule
Flexible, queue-based

Four principles of rating

A “verified” conclusion is only worth something if the measurement isn't biased.

  • Blind

    You see only an attempt code, the task, the rubric and the recording. No name, email, organisation, teacher or AI score before you submit.

  • Independent

    You are never assigned learners you teach or taught in the last 90 days, nor items you authored.

  • Calibrated

    You rate a benchmark set with reference scores before live work, and periodically after, to keep the scale stable.

  • Adjudicated

    When two ratings disagree, the attempt goes to an adjudicator. Every rating is kept, never overwritten.

Assessor track

Permissions open up step by step based on calibration results.

  1. 1

    Rater

    Rate attempts assigned in your queue, against the rubric.

  2. 2

    Calibrated rater

    Rate calibration sets that keep the shared scale in line.

  3. 3

    Adjudicator

    Decide the final score when ratings disagree.

Who we're looking for

  • Proficient in the language being rated (native or C1 equivalent and above)
  • Experience teaching, examining or working in that language
  • Meticulous, consistent, open to feedback on rating drift
  • A computer, good headphones and a stable connection

Frequently asked questions

Will I know whose attempt I'm rating?

No. You see only a pseudonymous attempt code, the task, the rubric and the recording. Name, email, organisation, teacher and AI score are hidden until after you submit.

I teach on the platform — can I also be an assessor?

Yes, if you pass calibration. You will never be assigned learners you teach or taught in the last 90 days.

How does calibration work?

You rate a benchmark set with reference scores. If your deviation is within the allowed threshold you start receiving live attempts; calibration repeats periodically.

Which languages are you recruiting for?

English for Vietnamese speakers comes first, including target listeners from markets such as Singapore. Other language pairs will open gradually.

Apply as an assessor

Leave your details and we'll email you about your first calibration set.

Your details

Used only to contact you about rating work.

For example: English (native), Vietnamese

0/2000

We only use this information to reply to you. See the Privacy Policy.