The Skills Market Today: Data Engineering Is the Public Lead
53 new roles in today's crawl. After 7-day category cooldown, the public headline is Data Engineering (3 roles) - pipelines and proof you can show, not another soft title.
Today's Job Index is a rotation day, not a volume-chase day. The crawl refreshed with 53 new roles and 43 publishable atomic skills across 106 roles in the window. Cybersecurity and OKRs still appear often, but the 7-day lead cooldown keeps those categories off the public headline so a different buyer signal can surface.
Public lead: Data Engineering#
| Publishable skill | Roles in window | Notes |
|---|---|---|
| Data Engineering | 3 | Public lead after category rotation |
| Cybersecurity | 15 | Loud volume; on cooldown for public lead |
| OKRs | 11 | Business-ops volume context |
| SQL | 8 | Same data category family; volume context |
| Consulting | 5 | Skill volume context |
| CSS | 5 | Design volume context |
| Python | 5 | Language volume context |
Data Engineering owns the public lead with 3 roles. That is a proof-first capability: moving data reliably, making it queryable, and leaving artifacts a buyer can evaluate - not a synonym for "technical."
SQL shows up in 8 roles in the same window. Treat that as related volume inside the data family, not as today's public headline. We rotate the public lead so one Job Index category does not own the email and blog for more than a few days.
These are counts from inside the crawl window, not a share of the whole market. Do not invent week-over-week market-wide coverage percentages.
What Data Engineering demand is really buying#
When postings cluster on data engineering (and adjacent SQL work), buyers are usually screening for three proof pieces:
- A pipeline or model you can show - not a resume line that says "owned data."
- Reliability under change - schema drift, late sources, and the decisions you made when the feed broke.
- Decision support - data that actually changed a weekly operating call, not a dashboard graveyard.
If you sell fractional analytics, platform, or ops work, that triad is the artifact pack: one real pipeline diagram or SQL notebook, one incident note, one before/after decision story.
How to use today's signal (90 minutes)#
If you are hiring: stop filtering by title overlap. Ask for a redacted pipeline sketch, the failure mode they fixed last quarter, and how the output changed a decision. Score the demonstration, not the last job title.
If you are building a practice: ship one public artifact this week - a short teardown of a messy data path you cleaned up (or a synthetic case with honest constraints). Attach it where buyers already look for operators.
If you are job searching: map your last three wins onto input → transform → decision evidence. Put that above the title line on your outreach.
Close#
Proof compounds faster than title updates. Rocket is built for that path: introductions and repeat work when your capability is visible, with Ember available to practice the conversations that convert.
Sources#
- Job Index export: job-index-2026-07-31.json (snapshot 2026-07-31, VPS operating-system Job Index).
- Public lead from lead-skill rotation (data_engineering); volume board for context only.
- Use publishable skills for public output; broad soft-skill tags retained for audit only.
- Do not invent week-over-week deltas or market-wide coverage percentages.