Use case
Run a campus drive without a spreadsheet in the middle
Volume hiring does not think in invitations, it thinks in cohorts: one named event, two thousand students, a four-hour window on a Saturday. ProctoHire has that as a first-class object, with a live console that answers the only question that matters on the morning — how many turned up.
- Cohort object
- Roster + window
- One assessment, one time window, one named event.
- Console counts
- Derived
- Computed from the data on every load, never a stored counter that can drift.
- Email is optional
- A mode
- Issue links without sending anything, when the placement cell distributes them.
- Scale plan
- 5,000 / month
- Assessments started per month, 15,000 candidates, 75 seats.
A drive is a noun, not a batch job
Most platforms let you invite a thousand people one thousand times. That is not the same thing, and the difference shows up at 09:15 on the day.
A drive owns four things: who is in the cohort, when they may sit it, which assessment they sit, and what state each of them is in right now. Because it is a real object, it has a name you can say out loud — the campus, the year, the programme — and everything that happens under it stays attached to it: the roster, the links, the sittings, the findings, the results.
The roster
Uploaded in bulk. Duplicates within an upload and against the existing roster are handled rather than silently creating two of the same person.
The window
A start and an end, honoured in your organisation’s own timezone rather than the server’s — set once in localisation settings and every deadline, expiry and printed date in the product follows it.
One assessment
The same paper for the cohort, so the results are comparable. Different cohorts are different drives.
A link expiry
Set per drive in hours, so a link handed out on Thursday is not still live in March.
The console during the window
Six numbers, all derived from the underlying data on every load rather than incremented into a counter that drifts by lunchtime.
The roster underneath is paginated at fifty rows — one screen — and each page resolves its sittings and invitations in a fixed number of queries rather than one per student. That sounds like an implementation detail and it is the reason the console is usable at two thousand names instead of timing out at four hundred.
Issuing two thousand links, and sending none of them
These are two separate operations in this product, on purpose.
Minting the invitations is fast and synchronous: you press the button, the links exist, and you can see them. Sending email is not batchable — it is one conversation per recipient — so it runs as a background job with a progress row the console polls, and you are not staring at a spinner wondering whether it worked.
And sending is optional. Plenty of placement cells distribute links through their own portal, a WhatsApp group or a printed sheet at the door, so "issue links, send nothing" is a first-class mode rather than something to work around. Where email is sent, it goes through your own configured sender, so it arrives from your domain and not from a vendor's.
Reviewing at volume without reviewing everything
Two thousand sittings produce more integrity signal than any human will read. The design question is not how to detect more, it is what to put in front of a person.
- Noise is worth zero and collapses by default. An unfocused window, a recording gap, a camera showing nothing — recorded, visible, scored at nothing, and folded away so they cannot bury the one finding that matters. One real sitting elsewhere in this product produced twenty clips of which eighteen were a window losing focus.
- The score is a queue order, not a verdict. It exists so a reviewer knows which sittings to open first out of two thousand. Every point of it traces to a finding you can watch.
- Similarity runs across the cohort. Two students in the same hall submitting the same answer is the characteristic campus problem, and it is exactly the case a paste detector is blind to, because both of them typed it. Answers are compared as normalised token streams, with the question’s own template and your corpus’s universal boilerplate struck out first.
- A scan writes nothing until someone looks. A similarity finding is an accusation about two named students, so the default is a dry run that persists nothing; recording it is a separate, audited, administrator-only act.
The detail is on proctoring and integrity evidence, and how a flagged sitting is actually reviewed is on session replay.
Fairness on a bad day
Campus infrastructure is not enterprise infrastructure. A platform that treats every disconnection as suspicious will fail an entire lab and blame the students.
A dropped connection is not a violation
Answers are saved locally as the candidate works and synced to the server, and a resume reconciles the two copies by which is newer rather than by which is convenient. Where a reload left a stretch unrecorded, that gap is written down and shown to the reviewer — and it is worth zero points, because a crash, a closed lid, an OS update and a deliberate escape are indistinguishable from a gap alone.
A second device is a finding, not a failure
Resuming the same invitation from a different machine is detected on the server and graded major but not fatal, because the honest cases are real — a dead laptop, a lab machine that rebooted, a student moving to a friend’s desk.
Adjustments are recorded, not improvised
Extra time, rest breaks, screen-reader mode and integrity exemptions are granted per candidate with a written reason. A grant cannot be deleted, only revoked with its own reason, and every change appends a row naming who did it. On a campus drive that record is what answers a complaint three months later.
The candidate’s clock, not yours
Windows, expiries and printed dates follow the timezone, date order and number format your organisation set, which matters the moment a drive spans more than one country.
What it will not do for you
- It does not invigilate a physical hall. This is a browser-based proctor. If the cohort sits in a lab under supervision, the supervision is still yours and the recording is corroboration rather than a replacement.
- It does not verify identity at the door by default. There is an optional photo-ID check at the start of a sitting, off unless your account owner switches it on, and it never scores against a candidate on its own.
- It cannot see a phone under the desk, a printed sheet, or the person sitting out of shot. Nothing that runs in a browser can.
- It will not choose your cut-off. The console tells you how many turned up, how many finished and how many produced findings. What score is a pass, and what a finding means for a nineteen-year-old, are decisions the product deliberately leaves with you.
- There is no per-institution portal for students to track applications across employers. A drive is your event, not a job board.
Elsewhere on this site
Remote and distributed hiring
One-way video screening, live interviews in the browser, and scheduling that works in the candidate’s clock rather than yours.
Security and data protection
Tenant separation, retention, erasure, sub-processors — and the certifications this product does not hold.
Pricing
The plan ladder, what one assessment means as a billing unit, and how a plan actually gets set today.
Buyer FAQ
GDPR, retention, appeals, dropped connections, and what happens when the evidence is inconclusive.
Run one drive with fifty names before you run one with two thousand
Build the paper, upload a small roster, set a window, and issue links without sending any email. You will know within an hour whether the console tells you what you need on the morning, which is the only question worth answering before a real drive.