what's actually happening in ai hiring right now
Software engineering job listings are up 30% in 2026. Over 142,000 tech workers have been laid off this year. Both of those things are true, and they're not a contradiction. They're the whole story. The market hasn't collapsed. It's split in two, and which side you land on depends almost entirely on what you do in the next 90 days.
what's actually happening
the layoffs are real, and the reason matters.
According to layoffs.fyi, over 113,000 tech workers had been laid off through mid-May 2026 across 179 companies, averaging 825 per day. About 48% of tracked layoff events have been explicitly attributed to AI, automation, or machine learning. But here's what's missing from most coverage: this isn't the 2023 correction. Back then, companies cut because they overhired. In 2026, companies are cutting profitable, growing businesses to fund AI infrastructure. Meta alone is spending $125–145 billion in AI capex this year, four to five times what they spend on their entire human workforce. The layoffs aren't a cost-cutting story. They are a capital reallocation story, and that distinction matters for how you position yourself.
the split: who's getting hired and who isn't.
AI/ML engineer postings are up 85% year over year as of mid-2026. General software engineering postings are down 49% from pre-pandemic peaks. That is not a gradual shift. These are two completely different markets sharing the same job board. Roles with at least two AI skills listed pay 43% more than comparable roles without them, and AI/ML salaries are up 20–30% year over year, the only segment in tech seeing real comp growth. Average AI engineer comp sits at $228K per AIDevBoard's analysis of 8,931 active listings.
the junior market is getting hit hardest.
Stanford's digital economy lab found that employment for software developers aged 22–25 fell nearly 20% from its 2022 peak, while developers over 26 saw headcount grow 6–12% in the same window. The mechanism: AI tools now handle the boilerplate coding, routine testing, and scripted debugging that junior roles were historically built around. Entry-level postings fell from 8.1% to 7.4% of the total IT job mix year over year, while senior-level postings climbed from 38.8% to 43.1%. If you're early in your career, the standard playbook doesn't work anymore. The good news is there's a better one.
where the actual opportunity is
The biggest mistake engineers make right now is applying to the market that used to exist. Big tech is doing strategic hires, not growth hiring, and the competition for those seats is brutal. The real opportunity is in three places:
AI-native startups building on top of foundation models. Companies like Cursor, Perplexity, Anduril, Anthropic, and OpenAI are hiring aggressively. These aren't traditional startup bets. AI-native startups are reaching product-market fit in 12–18 months, not 5–7 years, and starting comp at well-funded ones ($160K–$180K base) is competitive with traditional SaaS.
Mid-market SaaS companies adding AI to existing products. These companies need engineers who can integrate LLM APIs, build RAG pipelines, and ship AI features into real products. They're competing for talent against the labs but have far less competition from candidates.
Non-tech industries building AI teams from scratch. Healthcare, finance, legal, and logistics: these companies are standing up dedicated AI teams and they're not on most candidates' radar yet. That's the point. Domain knowledge plus AI skills is the combination that wins there. Small businesses are expected to hire nearly 1 million graduates this year, specifically targeting people who grew up with AI as a native tool. As economist Aaron Terrazas at Gusto put it: "Large companies are playing defense. Small businesses are playing offense."
the playbook: what to actually do
build something with AI, not just a knowledge of AI.
Candidates who can demo working LLM applications consistently outperform peers with equivalent credentials. "I've used ChatGPT" isn't a skill. Building a RAG pipeline, deploying an agent, or integrating an LLM API into a real project is. Employers want to see repos and deployed work, not coursework. The skills commanding the biggest premiums right now:
Skill | Why it commands a premium | Where to build it | Who's hiring |
LLM fine-tuning | Foundation of most AI product workflows; scarce at production scale | Hugging Face, fast.ai, personal projects with open-source models | AI-native startups, frontier labs, enterprise AI teams |
Retrieval-augmented generation (RAG) | The dominant architecture for enterprise AI right now; every company building on LLMs needs it | LangChain docs, open-source RAG projects, internal tools | Mid-market SaaS, healthcare, legal, finance |
Agentic AI frameworks (LangChain, LangGraph) | Fastest-growing subcategory; agentic AI job postings up 280% YoY | LangChain documentation, agent tutorials, ship one real agent | AI-native startups, Cursor, Perplexity, Anduril |
MLOps & model deployment | Chronically undersupplied; hundreds of listings, few qualified candidates | AWS SageMaker, Azure ML, Kubernetes for inference workloads | Any company running AI in production |
Python (AI stack) | Required in 71% of AI job postings; table stakes but still a filter | PyTorch, TensorFlow, scikit-learn projects; visible GitHub repos | Across all segments, junior to senior |
AWS / Azure cloud deployment | Cloud deployment skills are as crucial as ML knowledge for most production roles | AWS Certified ML Specialty or Google Professional ML Engineer (20–25% salary premium) | Enterprise AI, non-tech industries building first AI teams |
rewrite three LinkedIn bullets before you change anything else.
If your profile still describes feature engineering or hyperparameter tuning, you're being algorithmically deprioritized against candidates whose recent work describes "production agent evaluation" or "agentic workflows." The keyword drift in LinkedIn's hiring algorithm is real. Rewrite three bullets to describe your AI-integrated work before you touch anything else.
stop filtering out companies outside of big tech.
This is the non-obvious move in summer 2026. Nearly 974,000 graduates will be hired by small businesses this year, and those companies are specifically recruiting people who are native to AI tools. IBM announced plans to triple entry-level hiring in 2026. Non-tech companies in finance, healthcare, and logistics building their first dedicated AI teams have far less candidate competition and often offer more ownership over the actual work.
position yourself as an operator, not a learner.
The candidates landing roles are the ones who can talk honestly about something they tried, what didn't work, and what they learned. Not 'I'm learning prompt engineering' but rather 'I built this thing, it failed in this specific way, and here's what I figured out.' Context engineering is the emerging differentiator. Companies don't just need people who can use AI, they need people who can design how AI fits into a workflow, catch hallucinations, and make judgment calls on outputs. That's a human skill AI can't replicate.
target roles where AI skills are chronically undersupplied.
MLOps, Kubernetes, and distributed systems appear in hundreds of listings with a fraction of the qualified applicants. Full-stack AI engineers, people who can build the product around the model, not just the model itself, are exactly what mid-market companies need and can't find. If you're a full-stack engineer who can add solid AI tooling knowledge, that combination is landing roles right now.
the takeaway
This market rewards people who move fast and get specific. The engineers winning in June 2026 are treating AI fluency as a daily operating layer, not a line item on a resume. The job market hasn't collapsed. The eligibility criteria changed. Get current on those criteria and the door is open.
