11. Special datasets

Hiring signals (JOBS, HIRE)

JOBS reads one company's open roles over time from its own careers site, and HIRE compares hiring momentum across companies.

  • 4 min
  • 3 questions
  • Lesson 4 of 8

Why you would care

Where is this company hiring, and is it speeding up or slowing down? Hiring is a leading indicator: a company builds teams before it launches products or enters markets, and freezes hiring before it cuts costs. A surge in AI roles, a new office abroad, or a sudden halt in postings can show strategy months before an earnings call. It is one of the classic "alternative data" signals used by hedge funds.

The idea from scratch

An open role (requisition) is a job posting on a company's careers site. Counting them over time gives:

  • Open roles

    how many postings are live.

  • Additions / removals

    new postings and closed ones each week.

  • Net change

    additions minus removals.

  • Mix

    by function (engineering, sales, finance...), location, seniority, topic (AI, robotics...).

Limits:

  • Open roles are not hires, employees or headcount.
  • A closed posting may be filled, cancelled or just reposted.
  • Coverage depends on the careers systems that can be read.

A z-score of the weekly change (how unusual it is against the company's own past) flags surges and freezes.

Diagram: Company careers site leads to Open roles snapshot, weekly; Open roles snapshot, weekly leads to Additions, removals, net change; Open roles snapshot, weekly leads to Mix: function, location, seniority, topic; Additions, removals, net change leads to Unusual vs own history: surge or freeze.

See it in Gloom

Gloom screenshot: JOBS NVDA: open roles, postings in the last 30 days, new this week, remote share, median posting age, the open-roles trend, topic chips and the role list
JOBS NVDA: open roles, postings in the last 30 days, new this week, remote share, median posting age, the open-roles trend, topic chips and the role list. Historical example.

open themJOBS NVDA for one company; HIRE for the board across covered companies, HIRE NVDA for one. Tabs in JOBS: Roles, Locations, Seniority, Pay.

  1. 2,653 open roles, 1,071 posted last 30d, 170 new this week, 23% remote, 38d median age.
  2. The chart: open roles peaked on Sep 21 and dropped on Sep 22 to 23, then recovered.
  3. Topic chips: AI / ML 1369, Semiconductors 1197, Autonomy 575, Robotics 471, Data centers: where the company is hiring.
  4. The list: role, function, location, posting age (1d).
  5. Locations such as Taiwan, China and India: where engineering capacity is being built.

HIRE adds weekly additions, removals, net change and z-scores, signals read in tone (a surge green, a freeze red), role-family mix and peers.

Practice and recap

Try it3 tasks
  1. What share of NVIDIA's open roles in the chips mention AI / ML? (1,369 of 2,653: about half.)
  2. Why is "170 new this week" more informative than total open roles? (It shows current activity, not old postings still open.)
  3. Open roles fall 10% in a week. Name two explanations other than a freeze. (A batch of roles filled or closed at once; postings moved to another careers site.)
Common mistakes4 mistakes
  • Reading open roles as headcount.
  • Over-reacting to one week's change.
  • Comparing companies with very different posting habits.
  • Ignoring location and function shifts, often the most useful part.
Check yourself3 questions
  1. What is an open role?
  2. What does net change mean?
  3. Name one limit of hiring data.
Answers
  1. A job posting live on a company's careers site.
  2. New postings minus removed postings over the period.
  3. Any one: postings are not hires; closed postings may be cancelled or reposted; coverage depends on readable careers systems.
Words in this lesson5 words
alternative data
Non-financial data that hints at business activity.
open role (requisition)
A live job posting.
additions / removals
New / closed postings in a period.
hiring freeze
A pause in new hiring.
mix
The split by function, location, seniority or topic.

Educational material about reading market data, not investment advice.