We analyzed 3,502 open roles at 16 top tech companies. Here's what the data says.

We built a nightly pipeline that scrapes public job feeds from Greenhouse, Ashby, and Lever. Every morning it pulls fresh listings from 16 of the most closely watched employers in tech.

Yesterday we published the first snapshot: The AI-Era Hiring Index, 2026 Q2. It covers 3,502 open roles across Airbnb, Anthropic, Coinbase, Datadog, Discord, DoorDash, Figma, Instacart, Notion, Pinterest, Plaid, Ramp, Scale AI, Spotify, Stripe, and Vercel.

This is not a survey. It is not a sentiment index. It is a direct count of what these companies are actually hiring for right now, extracted from their own applicant tracking systems.

Here are the findings that stood out.

Salary transparency is at 44.9%, but the variance is enormous

Across the full index, 44.9% of job descriptions disclose a salary range directly in the listing. That headline number hides a massive spread between companies.

Discord leads with 98.7% of its listings including salary ranges. Pinterest is at 82.4%. Anthropic discloses in 71.9% of roles. At the other end, Spotify shows 0% salary disclosure in its job descriptions (their salary information is shared through other channels, not in the JD itself).

The most striking number belongs to Anthropic. Across 307 roles with disclosed ranges, the median salary band is $290,000 to $365,000. That is not a cherry-picked senior engineering number. That is the median across all disclosed roles at the company.

For candidates, the takeaway is clear: research the specific company's disclosure practices before you negotiate. The information landscape varies dramatically even among employers of similar size and prestige.

Remote work signals are unreliable (and we can prove it)

We measured remote work two ways: by the structured ATS flag that companies set on each listing, and by scanning the actual job description text for remote work language. The results diverge significantly.

ATS flags mark 24.5% of roles as remote. Text-based scanning finds remote work mentioned in 44% of listings. Both numbers are "correct" in a technical sense. Neither tells the full story.

The divergence is sharpest at individual companies. Figma flags 0% of its roles as remote in structured data, but 100% of its job descriptions mention remote work. Ramp flags 97.7% as remote, but only 4.6% of its JD text mentions it. Notion flags 62.6% remote while only 2.6% of its descriptions reference it.

We publish both signals because neither is authoritative on its own. If you are filtering job boards by "remote" tags, you are seeing a filtered and potentially misleading picture. Read the actual job description.

LLM experience is no longer a specialty

The top skills data tells a story about how fast AI has moved from a niche to a baseline expectation.

Across engineering roles (891 listings), the top mentioned skills are:

1. Python (33.4%)
2. AWS (21.9%)
3. LLM / foundation model experience (21.2%)
4. Go (19.2%)
5. Java (19.1%)

LLM experience at 21.2% of engineering roles would have been unthinkable two years ago. But the real signal is in specialized roles. In data and ML positions (333 listings), LLM experience appears in 56.5% of job descriptions. In research roles (69 listings), it is 65.2%.

At these companies, working with large language models is no longer something that differentiates your resume. It is something employers assume you can do.

Engineering leads, but sales is bigger than you think

The role mix across 3,502 positions:

• Engineering: 891 roles (25.4%)
• Sales/GTM: 624 roles (17.8%)
• Data/ML: 333 roles (9.5%)
• Operations: 257 roles (7.3%)
• Marketing: 161 roles (4.6%)
• Finance/Legal: 146 roles (4.2%)
• Product: 132 roles (3.8%)
• Research: 69 roles (2.0%)
• Support: 51 roles (1.5%)
• People: 29 roles (0.8%)
• Design: 27 roles (0.8%)

Engineering dominance is expected. What is less expected is that sales and go-to-market roles account for nearly 18% of all open positions. These are companies with strong products that still invest heavily in distribution. For job seekers outside of engineering, that is a meaningful signal about where the opportunities are.

Why we are publishing this

Most hiring data you encounter is secondhand. Surveys with self-reported numbers. Aggregations with opaque methodology. Reports behind paywalls that you cannot verify.

We wanted to do the opposite. The full dataset (3,502 rows) is published under a CC BY 4.0 license. You can download it, analyze it, and use it for your own research. Our methodology documents exactly how we classify roles, extract salaries, detect remote signals, and identify skills. Every headline number in this post is reproducible from the raw data.

If you are a journalist, analyst, or researcher covering labor markets, you can cite this data directly. If you are a candidate, you can use it to benchmark your target companies and roles against the broader index.

What is next

This is Issue 01 of a quarterly series. The Q3 snapshot will capture how these numbers shift as the year progresses. Over time, the index will show whether salary transparency is increasing, which skills are gaining or losing demand, and how the role mix at AI-era companies evolves.

Read the full report →

Download the dataset →

Review our methodology →