Last week a staff engineer I know got an offer from a Series C company after a 14-week search. He had 11 years of experience, a strong referral, and a good GitHub profile. Three of the companies he applied to never responded at all. He took the offer, which was 18% below his last total comp. He considered that a win given the timeline.
That’s the 2026 tech job market in a single data point: things are better than 2024, and they’re still not good. Both of those are true at the same time.
What the actual data says
The LinkedIn Economic Graph reports tech job postings recovered meaningfully through late 2025 and into early 2026, with software roles up roughly 28% from 2024 lows. That sounds like good news. The context that matters: applications per posting also went up, because the pool of candidates grew faster than the postings did. More jobs, but more competition per job.
The Stack Overflow Developer Survey 2024 found that 62% of developers who changed jobs did so because of compensation concerns, not because they were laid off. That’s important because it tells you the market isn’t purely about displaced engineers looking for work. A lot of the competition is employed engineers looking for better situations. They’re a harder baseline to beat.
Which sectors are actually hiring
It’s not uniform. Some of this is obvious. AI infrastructure companies, enterprise software companies adding AI capabilities to existing products, and healthcare tech companies dealing with a multi-year backlog of digitization work are all hiring meaningfully. Defense and government contractors have had steady demand that doesn’t correlate much with the broader tech cycle.
FAANG is more complicated. Meta has been selectively hiring for specific teams while staying flat overall. Google has been cutting some teams while growing others, which is a different thing than growing. Amazon has been hiring in AWS more than in retail tech. The headline numbers for any of these companies often obscure what’s actually happening at the team level.
Where I’d be skeptical: early-stage startups (Series A and earlier) are having a genuinely hard time competing on compensation right now, and many of them are running lean and staying lean until their funding environment improves. If you’re offered a pre-seed role with below-market salary and equity as a primary component, that’s a different kind of bet than it was in 2021.
The skills that changed the fastest
Probably not surprising: anything adjacent to LLMs and AI systems is commanding a premium. But there are two things that are true here that are somewhat in tension.
First, every company wants engineers who understand how to build with AI. Second, there are not many engineers who have real, substantial production experience building AI-native systems. The gap between what’s in job descriptions and what candidates actually have is large. Companies are frequently asking for things that only exist in a few thousand engineers globally, then waiting six months to fill the role, then adjusting the requirements. I’ve seen this cycle happen enough times that it’s a pattern now.
Platform engineering, security engineering, and full-stack development with AI capabilities are all areas where I think the demand-to-supply ratio is favorable for candidates. “AI/ML engineer” as a job title is competitive because everyone is applying. “Infrastructure engineer who understands ML deployment pipelines” is a narrower search with a smaller candidate pool.
Compensation: the honest picture
Total compensation at major tech companies has come down from 2021-2022 peaks. Not catastrophically, but noticeably. Senior software engineers at FAANG-tier companies who were at $400K total comp in 2022 are more likely to be at $320K to $360K for comparable roles in 2026, partly because stock valuations normalized and partly because the balance of power has shifted back toward companies in this market.
Entry and mid-level ranges have been more stable in absolute terms but are below inflation-adjusted peaks. Whether or not that matters depends heavily on where you live and whether the company is offering remote work. A $150K role in Austin is a different financial situation than a $150K role in San Francisco.
I should be clear: I don’t have precise, sourced data on compensation ranges across the industry. These are patterns based on reported data from Levels.fyi, LinkedIn salary insights, and conversations. Your specific situation will vary.
What’s actually different about interviewing now
The interview process has compressed at a lot of companies. Where a full loop used to take six weeks, many companies are now trying to close in three weeks or less. This is partly candidate experience improvements and partly the fact that good candidates are fielding multiple offers simultaneously and slower processes lose them.
The behavioral component has gotten more weight at senior levels. After several years of hiring engineers primarily on algorithmic coding performance and then watching those engineers struggle with ambiguity and cross-functional work, a lot of hiring managers have rebalanced. Expect more time on “tell me about a time you disagreed with a technical direction and what you did about it” than on “reverse a linked list.”
AI integration questions are common now. Companies want to know whether you’ve actually used AI tools in your development workflow, and how you think about when to use them versus when not to. Candidates who have a practiced, specific answer here are doing better than candidates who give a vague “yes, I use Copilot sometimes.”
The referral situation hasn’t changed
Referrals are still how a disproportionate fraction of tech hires happen. If you have a warm connection at a company you want to work at, that path is almost always faster and more likely to result in an interview than a cold application. This isn’t a secret. It’s just inconvenient and uncomfortable for people who aren’t well-networked, and they ignore it in favor of more applications.
I don’t know if the referral advantage has grown or stayed stable over the past few years. What I do know is that it hasn’t gone away. If you’re doing a job search and your primary activity is submitting applications on LinkedIn and Indeed, you’re doing the part of the job search with the lowest return on time invested.
The 2026 market is selective, not dead. The engineers I’ve watched navigate it successfully had one thing in common: they ran a structured, targeted search rather than a high-volume spray-and-pray campaign. Fewer applications, more research per application, more conversations before the formal process starts. That pattern works better in a selective market than in a hot one. This is a selective market.