Generated by Codex with GPT-5

The Pragmatic Engineer surfaced this May 26, 2026 deepdive through its latest archive and feed, and the original piece is State of the software engineering job market in 2026.

The useful thing about this report is that it refuses the two simplest stories about engineering hiring. The market is not back to the easy-growth world of 2021, but it is also not frozen. Hiring is recovering, unevenly, and the shape of that recovery says a lot about what companies now value: proven engineering ability, AI-adjacent skills, industry context, and proximity to the places where companies are still willing to concentrate teams.

The Pragmatic Engineer and Jessica Salmon built the piece around data from TrueUp and Workforce.ai. TrueUp tracks open roles at Big Tech, leading startups, and scaleups that tend to pay in the top compensation tiers. Workforce.ai tracks large-scale employment movement across companies, roles, levels, industries, and locations. That combination makes the article more useful than an anecdotal job-market mood check. It compares open roles, actual headcount changes, geography, and company-by-company momentum.

A Recovery With Limits

The headline is cautiously positive. Top tech companies are listing more software engineering jobs than a year ago, and TrueUp’s data shows a steady increase in openings since March 2023. The article says the highest-paying tier of companies is hiring about 20% more than last year.

But the recovery is not broad enough to feel like a boom. Indeed data cited in the piece shows US software developer listings still below May 2021 levels. Over the past year, the US and UK are up, Canada is roughly flat, and Germany and France are down. The authors read that as a sign that US-headquartered companies are reopening hiring faster than many European-headquartered employers.

The Workforce.ai view adds another constraint: hiring is seasonal and fragile. Most net growth happens from March through June, after annual headcount budgets are set. The second half of the year has been much weaker, and the data shows recent third-quarter contractions tied to layoff waves. That matters because it makes the recovery easy to overstate. Hiring can look healthy in spring and still feel harsh by late summer if companies pull back again.

The piece also frames 2023 as unusually bad for software engineers. It was the worst period in two decades by the article’s reading of the data, with the industry shrinking rather than merely slowing. The current rebound should be understood against that baseline: better than the trough, not necessarily better than the long-term pre-correction market.

Employer Stability Is Diverging

One of the sharpest sections compares Big Tech headcount changes since May 2024. Meta is the outlier. It grew software engineering headcount by nearly 20% over two years, then cut 10% of staff and reassigned thousands of software engineers to manual data-labeling work. The article describes Meta as having moved from one of the most stable places to work in tech before 2022 to one of the riskiest in job-security terms.

The other large tech companies look much steadier, though not necessarily fast-growing. Apple grew software engineering headcount by about 10%, Google by 5%, while Microsoft and Amazon were slightly down. The article’s implication is that “Big Tech” is no longer one uniform labor market. Apple and Google look relatively stable. Amazon and Microsoft look disciplined. Meta looks volatile.

Public tech companies tell a similar story of slower but still meaningful growth. Shopify, Stripe, Atlassian, Snap, and Spotify grew over the past two years, but the most recent year is less aggressive. Stripe still stands out, growing software engineering headcount faster than nearly all of Big Tech. Atlassian and Shopify also expanded faster over two years than Google, Microsoft, or Amazon, though the article notes that recent layoffs at Atlassian and Snap were not yet captured in parts of the data.

For engineers, this is a reminder that company quality and company stability are no longer the same question. A prestigious employer can still whipsaw teams. A less culturally central employer can be the one still hiring. The safer move is to evaluate the business cycle of a company, not just its brand.

Where The Openings Are

The companies with the most software engineering openings are familiar: Apple, IBM, and Amazon remain the top three. The more interesting part is the turnover around them. Accenture, Tesla, Cadence, HPE, and SpaceX entered the top 20. Meta and Oracle dropped out, with Oracle having announced up to 30,000 layoffs in March and Meta cutting after its rapid hiring cycle.

Google is one of the largest companies advertising more roles than last year, with 62% more engineering openings listed. Hardware companies also show up as stronger software employers: Micron, Qualcomm, and AMD all have more software engineering openings. That is an important signal because it connects the software labor market to the AI infrastructure buildout. Chips, devices, data-center systems, and tooling all need software talent.

Geographically, the top roles remain concentrated. Most of the jobs tracked by TrueUp are in the US, followed by India, the UK, and Canada. That concentration fits the rest of the report: the market is improving, but the easiest paths are still clustered around the employers, hubs, and categories with budget.

The fastest-growing companies make the market’s direction even clearer. Ramp grew software engineering headcount by 94%, Wiz by 84%, Datadog by 68%, Rippling by 55%, Figma by 41%, and Netflix by 37% over the past two years. The categories are revealing: fintech, security, observability, design tooling, and broader entertainment infrastructure.

Datadog is especially telling. The article argues that AI agents are increasing demand for observability, and notes that OpenAI reportedly spent around \$170M on Datadog in 2025. If AI makes software systems more automated, distributed, and harder to reason about, then companies that help teams monitor and debug those systems become strategic infrastructure.

AI Is Becoming A Hiring Filter

The public portion of the article reaches the start of the AI engineering section, and the article’s own outline gives the broader point: AI engineering demand is rising fast, many tech companies are prioritizing AI engineering recruitment over general software engineering, and companies like Apple, Google, and TikTok have the most AI engineering openings. The outline also says many larger tech companies have 50-100% more AI engineering listings than a year ago.

The most important question is not whether AI engineering is replacing software engineering hiring in a clean one-for-one way. The article explicitly says the data does not prove that. The more practical interpretation is that AI engineering is becoming part of the expected toolkit for many software roles.

That distinction matters. A company may not label a role “AI engineer,” but still expect candidates to understand LLM APIs, retrieval, evals, agent workflows, model limitations, and how AI features fit into an existing product. The boundary between software engineering and AI engineering is getting blurrier. In that environment, engineers do not need to become researchers, but they do need to understand how to build reliable systems around models.

The Takeaway

The 2026 software engineering market is selective, not dead. It rewards engineers who can point to applied judgment, not just generic coding skill. It rewards people who can work in or near AI-heavy product areas, but it does not say that every engineer must become a model specialist. It rewards proximity to strong hiring markets, but it also rewards choosing a company whose business cycle is expanding rather than contracting.

The report is most useful as a correction to lazy narratives. AI has not made software engineers obsolete. Big Tech is not uniformly safe. Startups and public companies are not uniformly risky. Hiring is up, but uneven. The strongest opportunities are in companies building or selling the infrastructure that the AI cycle needs: security, observability, fintech automation, design tooling, hardware, and platforms that make complex systems manageable.

For an engineer deciding what to learn next, the practical advice is straightforward: keep the software fundamentals sharp, build enough AI engineering fluency to be credible, and develop domain knowledge in a market where companies are still spending. The hiring market is no longer paying equally for every generic engineering profile. It is paying for people who can help companies turn the AI cycle into reliable products and durable businesses.