Work that puts AI or machine learning into a live product is still the tech job to chase if pay and staying hireable both matter. Something shipped. Users actually touch it. A chatbot demo is not that, and neither is a prompt-only title with no owner.
Hiring is uneven. Indeed Hiring Lab found U.S. postings only about 6% above pre-pandemic levels late in 2025, while ads that mention AI jumped more than 130%. Their AI Tracker reached 4.2%. Almost 45% of data and analytics postings used AI language. Marketing and HR ran far smaller shares.
"Best" is a skills mix employers already budget for, not a trophy title. AI-heavy engineering sits at the top of posted pay. Cyber and cloud sit close behind, because outages and audits don't wait for a hiring thaw.
Do you need a computer science degree to get in? Plenty of people don't have one. They lean on certs, a public project, and job boards that filter by tools instead of school.
What U.S. tech hiring looks like right now
Thing is, the market can look hot and frozen in the same week. A 2026 hiring recap that works through CompTIA and Dice data expected only about 1.9% net tech employment growth, toward roughly 9.8 million workers.
Dice, using Lightcast in that same recap, had U.S. tech postings up 18% year over year in August. AI and machine learning ads were up 101%.
Software sits in an awkward middle. The recap's Indeed analysis said jobs with AI in the title drove 37% of software posting growth between May 2025 and May 2026. Senior roles made up 69.3% of software development postings in the first quarter. Recent computer engineering graduates showed a 7.8% unemployment rate in that write-up, second among majors.
Read that as a timing problem. It is not a closed door. Juniors still get hired. Just slower. Proof of shipped work matters more than a generic "passionate about tech" line.
AI tools are already normal on the job. The 2025 Stack Overflow survey, cited in later labor write-ups, had 84% of respondents using or planning to use AI in development, and about half of professional developers using it daily. That still leaves a person in the loop. Employers want someone who can debug what the model missed.
Roles worth targeting
Pay figures bounce by city, by level, and by whoever published the roundup. Treat them as signals, not an offer letter.
| Role | Best for | Pay and growth signals | Usual way in | Where listings cluster |
|---|---|---|---|---|
| AI / ML engineer or specialist | Highest posted pay | A third-party 2026 roundup put AI engineer averages near $206,000 after a 143.2% jump in 2025; staff ML and cloud architecture ads often sit in the $180,000-$220,000 band on recruiting boards | Python, a public model or retrieval app, cloud ML practice | Indeed, LinkedIn, company career pages |
| Information security analyst | Shortage and funded headcount | About $124,910 median and roughly 29% to 32% projected growth in federal figures those roundups cite | Security+, labs, help desk or SOC-adjacent work | Indeed, LinkedIn, USAJOBS for federal roles |
| Software developer | Volume of openings | Around $132,000-$133,000 median, with decade growth often cited in the 15% to 25% range depending on the BLS vintage | GitHub, internships, bootcamp or self-taught projects | LinkedIn, Indeed, employer sites |
| Data scientist / analytics | AI language already in the ads | About $112,590 in the same BLS-citing roundup; nearly 45% of Indeed analytics ads mention AI | SQL, Python, a public analysis notebook | LinkedIn, Indeed |
| Cloud, DevOps, network architect | Production systems that still break | Network architect about $129,840; cloud architecture often higher | AWS, Azure, or Google Cloud practice plus a deployed stack | LinkedIn, Indeed |
| IT support, QA, SOC tier 1 | First paycheck and a company login | Lower cash than specialist tracks; use them as a bridge | Vendor-neutral certs, ticket or test evidence | Indeed, LinkedIn, staffing firms |
IT managers show up near $171,200 in that same roundup. That's usually not an entry seat. Skip it until people and systems are already on your resume.
The AI and machine learning path
Posted pay sits at the top here, which is why people treat this as the one best tech job. Demand is real. Competition is loud, and that combo trips people up.
You'll see ML engineer, AI engineer, applied scientist, and data scientist treated as twins. They aren't. ML postings usually want train, evaluate, deploy. Data science postings lean on analysis, experiments, and explaining the result to a non-engineer. Read the first ten bullets of any ad. Apply to the one you can already do.
Build something ugly that works. A small retrieval app counts. So does a classifier on a public dataset, or a fine-tune where you wrote down what failed, and why, and what you'd try next time, even if the honest answer is "retrain with better data." Recruiters glance at public code for a few seconds. Put the README first.
Search like a specialist, not a tourist. On Indeed, pair the title with a stack: machine learning engineer and Python, or AI engineer and AWS. On LinkedIn, sort by newest. If a listing already shows a crowd, move on. When both links exist, apply on the company site.
Don't wait around for a junior ML req that also wants three years of production ML. Those are scarce. Adjacent titles work: data analyst who uses AI tools, ML ops, analytics engineer, backend engineer on an AI team. A 2025 NACE finding, cited in later hiring research, had 61% of employers saying they're not replacing entry-level jobs with AI. The squeeze is real. Blanket replacement isn't the whole story.
Cybersecurity as a faster specialist track
Security stays funded when other budgets shrink. Roundups that cite federal data still put information security analysts near a $124,910 median, with growth in the 29% to 32% range. Training vendors often repeat a figure of nearly half a million unfilled U.S. cybersecurity jobs. Take that as a shortage signal, not a personal offer.
The on-ramp here is more cert-shaped than the AI path. CompTIA Security+, or an equivalent vendor-neutral cert, still maps cleanly to ATS keywords. Then pick a lane.
- SOC analyst, GRC, or cloud security
- Labs you actually ran: a home SIEM, an incident narrative, or a cloud IAM hardening checklist
Search Indeed and LinkedIn for SOC analyst, security analyst, and GRC analyst. Apply within the first day or two. Skip anything that sounds like breaking into a system you don't own. Employers want people who can explain risk in plain English and keep the lights on.
Cloud, software, and data jobs
Cloud and DevOps work is the unglamorous cousin that still pays. When a cluster drops client requests, someone has to find the cause. Certificates from AWS, Azure, or Google Cloud help because they match words parsers already look for. Pair the cert with a deployed project, though. A cert alone is a PDF.
Software development remains the volume play. Later BLS-citing roundups still put medians in the low $130,000s with double-digit decade growth. The catch is seniority. If most software ads want seniors, your first title might be support engineer, QA, or intern-to-hire. That's a path, not a failure.
Data work is where the AI language is thickest. Nearly 45% of data and analytics postings on Indeed already talk about AI. That favors people who can write SQL, clean a messy table, and use a model without pretending they invented it. Prompt-style titles exist too, but pay bands on those ads swing wildly, and the work isn't standardized. If you chase one, look for a team that already owns a product. Not a listing that's only "write prompts."
First-paycheck roles for switchers
To be honest, a lot of career-changer advice skips the boring first job. Help desk, IT support, QA, and operations coordinator seats are how people get a company email and a system they can talk about in interviews.
You can burn an entire Saturday hopping from Indeed to LinkedIn to a company careers page, rewriting the same three bullets so the posting's keywords show up, and still feel like you applied into a void, especially when the listing says entry level and then asks for four years of Kubernetes. That's the junior squeeze. It isn't proof you should quit.
Use those first roles as paid labs. Ticket queues teach troubleshooting. QA teaches how products fail. Then you pivot the resume toward cyber, cloud, or AI with internal applications, which often beat cold applies from outside.
Gig-style data labeling can teach you how training sets work. It's a learning tactic, not the high-pay destination in the table.
How to search Indeed, LinkedIn, and company sites
Turns out spraying 80 generic applications in a weekend mostly trains software to ignore you. Several search write-ups land in the same place: 5 to 15 tailored applications beat a hundred copies, early applies get more eyes, and the company site sometimes ranks higher with recruiters than the one-click button.
- Pick one target role and two backup titles that share the same skills.
- Set Indeed and LinkedIn alerts for those titles plus one tool keyword (Python, Security+, AWS). Refresh the filters so you're not staring at cached junk.
- Limit results to the past 24 to 48 hours. Remote filters help, but nationwide remote tech is crowded and full of listings with no real company behind them. If the employer has no footprint, skip it.
- Rewrite the top of your resume to mirror the posting's tools. Keep it a simple file parsers can read. No text boxes. No icons standing in for job titles.
- Apply on the employer site when you can, then send a short LinkedIn note to a recruiter or hiring manager with one project link.
- Track what you sent. Drop boards that only return ghost jobs.
Don't treat any board's "jobs you might like" row as a ranking of the labor market. It's an engagement feed.
Skills, certs, and degrees
Many listings still mention a bachelor's. Many hiring teams will still talk to you if the resume shows shipped work.
Bootcamps and self-taught paths show up in developer origin stories all the time. They're not a shortcut around a portfolio. Certs help. They don't replace a project you can show. A cert gets you past some filters. The project gets you the call.
Remote listings and hours
Data and QA work usually beats on-call DevOps for sleep. If balance matters to you, search hybrid in your metro instead of worldwide remote.
FAQ
Which tech job pays the most without a degree? Specialist AI and ML seats sit at the top of posted pay in third-party roundups, sometimes near or above $200,000 at the experienced end. Cloud architecture and senior security follow. None of that is an entry offer. Certs plus a public project are the usual substitutes for the degree line.
Is software development still worth chasing? Yes, as a volume career, not as an easy junior market. Medians in recent BLS-citing roundups sit around $132,000-$133,000, but senior titles dominate a large share of ads. If you're new, plan a bridge title.
What skills show up most in 2026 postings? AI and machine learning language, cloud platforms, security, and plain troubleshooting. Indeed's tracker had AI terms in 4.2% of all U.S. ads by the end of 2025, and in nearly half of data and analytics ads.
Where should you actually apply? Indeed and LinkedIn for discovery, then the company site for the real application. Set alerts, apply in the first day or two, and keep the resume ATS-plain.
Pick one row from the table. Make two alerts today, one on Indeed and one on LinkedIn. Then send five tailored applications this week, each with a project link that matches the posting.