What are surprisingly high-paying jobs that most people have never heard of?
I’m curious about careers that pay really well but aren’t commonly talked about in school or online.
Not the typical doctor, lawyer, or software engineer.
What are some underrated or unknown jobs that have high salaries?
How did you discover them, and what does it take to get into that field?
Jobadvisor
It is easy to find information on surgeons or CEOs, but there is a whole "hidden" layer of the economy where people earn high six-figure salaries by doing jobs that are either too dirty, too dangerous, or too niche for the average person to consider.
Here are some surprisingly high-paying jobs that rarely make it into high school career brochures.
1. Underwater Welder
This is often cited as one of the most dangerous jobs in the world, which is exactly why it pays so well. These professionals repair oil rigs, bridges, and nuclear power plants deep underwater.
The Pay: Entry-level starts around $60,000, but "go-getters" and those working on deep-sea offshore rigs can earn $100,000 to $200,000+ per year.
What it takes: You need to be a certified commercial diver and a certified welder. It requires intense physical stamina and the ability to stay calm in pitch-black, pressurized environments.
2. Nuclear Reactor Operator
While most people think you need a PhD in Physics to work at a nuclear plant, many operators start with high school diplomas or military experience (especially from the Navy’s nuclear program).
The Pay: Average salaries often exceed $120,000, with senior roles hitting $160,000+ with overtime and bonuses.
What it takes: Extensive on-the-job training and a license from the Nuclear Regulatory Commission (NRC). It's a job of extreme boredom interrupted by moments of high-stakes precision.
3. Elevator Inspector / Mechanic
We use them every day, but rarely think about the people who keep them from falling. This is a highly unionized field with a very high barrier to entry due to the technical expertise required.
The Pay: Experienced mechanics often earn $80,000 to $120,000, while those in major cities with overtime can push $200,000.
What it takes: Usually a 4-to-5-year apprenticeship. It’s less about "fixing" things and more about complex electrical systems and safety compliance.
4. Ethical Hacker (Penetration Tester)
Not "unknown" in tech circles, but many people don't realize that companies will pay you a fortune to try and break into their systems.
The Pay: Mid-to-senior levels earn $130,000 to $180,000+.
What it takes: You don't necessarily need a degree, but you do need certifications (like OSCP or CEH) and a deep, "detective-like" curiosity for finding loopholes in code.
5. Air Traffic Controller
This is one of the few jobs where you can earn a "doctor’s salary" without a graduate degree, but it comes at the cost of extreme stress.
The Pay: The median salary is roughly $140,000, with many senior controllers earning closer to $200,000.
What it takes: You must start the FAA academy before age 31. The training is notoriously difficult, and the "washout" rate is high because you have to be able to track dozens of moving variables in 3D space simultaneously.
6. Flavorist (Flavor Chemist)
Ever wonder who designed the exact taste of a Flamin' Hot Cheeto? That’s a Flavorist. They sit at the intersection of organic chemistry and culinary art.
The Pay: Senior flavorists at major food labs can earn $100,000 to $150,000+.
What it takes: Usually a degree in Chemistry or Food Science, followed by a multi-year apprenticeship (often 7+ years) to be certified by the Society of Flavor Chemists.
Comparison at a Glance
| Job Title | Est. Salary Range | Education Required | Key Factor |
| Underwater Welder | $60k – $200k+ | Certifications | High Danger |
| Nuclear Operator | $90k – $160k | Training/Licensing | High Responsibility |
| Elevator Mechanic | $80k – $120k+ | Apprenticeship | Niche Skill |
| Air Traffic Controller | $120k – $190k | FAA Training | High Stress |
| Flavorist | $100k – $150k | Chemistry Degree | Sensory Talent |
How to discover more?
These jobs are usually found by looking at "Critical Infrastructure" or "Trade Monopolies"—industries that society cannot function without but that have high barriers to entry (like dangerous conditions, long apprenticeships, or strict federal licensing).
Is it no longer common to offer training to new hires?
I am in a People Ops role and I’ve raised my concern that we should give more attention to new hire training in our onboarding process. This means that documentation must be improved.
This is based on the feedback I received from exit interviews. We have had the issue of people leaving within 3-6 months and their main reason is that they felt like they never got proper training but were expected to deliver the same output as those who have been in the company longer.
Our founder said that this is a waste of time. If they can’t learn quickly, either they go or we fire them. He also said documentation is not a priority right now and nobody reads those anyway.
I feel surprised. I’ve been in bigger companies and training + documentation are pretty standard. Is this a new trend?
Jobadvisor
I can see why you’re surprised. Transitioning from a structured corporate environment to a founder-led startup (or a high-growth "hustle culture" environment) can feel like landing on a different planet.
To answer your question directly: No, it is not a "new trend" to abandon training. In fact, data consistently shows that structured onboarding increases employee retention by as much as 82%.
What you are experiencing isn't a trend; it's a Founder Fallacy. Your founder is likely operating under the "sink or swim" myth, which is often a symptom of short-term thinking over long-term scaling.
The Reality Check: Startups vs. Corporates
In larger companies, training is a risk-mitigation strategy. In smaller or founder-led companies, it's often viewed through these (misguided) lenses:
The "Rockstar" Myth: The belief that if you hire "A-Players," they should be magically clairvoyant and figure out the tech stack and culture by osmosis.
The Speed Trap: The founder feels they don't have time to document today, not realizing they are losing weeks of productivity tomorrow when a new hire is paralyzed by a lack of info.
The Cost of Churn: Founders often overlook the actual cost of losing someone in 3–6 months. Between recruiting fees, interviewing time, and lost momentum, it usually costs $1.5x$ to $2x$ that person’s salary to replace them.
Reframing the Conversation
Since your founder views training as a "waste of time," talking about "employee experience" or "learning journeys" might fall on deaf ears. You need to speak the language of efficiency and ROI.
| Founder's View | The People Ops Reality |
| "Documentation is a waste; nobody reads it." | Documentation is asynchronous training. It saves the existing team from answering the same questions 50 times. |
| "If they can't learn fast, fire them." | If they are talented but lack context, you are firing potential because of a process failure. |
| "Training is too slow." | Poor training leads to Rework. Doing a task wrong and fixing it takes 3x longer than doing it right the first time. |
Strategy: The "Trial" Documentation
If the founder won't back a massive overhaul, try a "Low-Lift" approach:
The "Loom" Method: Tell the founder and leads they don't have to write manuals. Just record their screen for 5 minutes while doing a task. That is the documentation.
The New Hire Wiki: Ask the next person who survives the 3-6 month mark to write down everything they wish they knew. It crowdsources the effort away from you.
The Cost Analysis: Present the data from those exit interviews. Show the founder exactly how much money is walking out the door every time a 4-month hire quits.
A note of empathy: It’s tough being the "voice of reason" in a room that values "move fast and break things." But remember, if the company keeps "breaking" its people, eventually there won't be a company left to run.
I used 9 AI models to rank 52 careers for automation risk and salary. Does this match what you're seeing in your industry?"
Why?
I was procrastinating on studying and got annoyed at all those "Top 10 Majors for 2026!" articles that are basically just SEO spam. So I decided to do something arguably dumber but more interesting: make a bunch of AI models fight it out over which college majors/careers are actually going to be viable in the next decade.
I used 9 different advanced AI models:
ChatGPT 5.2 (Thinking/Deep Research)
DeepSeek 3.2 (Deep Think)
Claude Sonnet 4.5 Extended
Gemini 3 Pro (Deep Research)
Grok 4.1 (Thinking/DeepSearch)
Perplexity Deep Research
Qwen3-Max (Deep Research)
Ernie 4.5 Turbo
Kimi k2.5 Agent Swarm
The setup:
I gave each model the same massive prompt (I'm simplifying it here, but the actual prompt was much more detailed). I made them score 52 specific college majors on a 100-point scale based on these factors:
Financial Core (40 points):
Salary trajectory, job volume, and how recession-proof it is
AI Survival Score (40 points):
Can you use AI to be 10x more productive, or does AI just replace you?
Does the job require physical skills or human empathy that AI/robots are terrible at?
Are there legal requirements that force a human to be involved? (medical licenses, PE stamps, CPA, etc.)
Human Factor (20 points):
Burnout vs. pay - is the suffering worth it?
How hard would it be for some random person with ChatGPT to fake your job?
Then I compared all 9 outputs and built a master tier list based on where they actually agreed.
S-TIER: The Safe Bets
These had near-universal agreement:
Nursing (NP/CRNA level)
AI can't do physical patient care and legally can't prescribe medications. The robots aren't there yet and won't be for a while.
Medicine (MD/DO)
Obviously. Licensing requirements are the ultimate moat. Brutal path but highest job security.
Electrical & Computer Engineering
AI can write software but can't physically design circuit boards or build chip infrastructure. Claude specifically mentioned the "semiconductor boom" - we need a decade+ of engineers just to build the data centers AI itself needs to run.
Cybersecurity
AI actually makes hackers more dangerous - they're using it to write exploits and malware faster than we can patch them. It's an arms race.
Warning though: Multiple models flagged that cybesec isnt entry level. You often need certs + experience just to break in now.
Specialized Engineering (Petroleum, Aerospace)
Super regulated and niche. Interestingly, Claude called Petroleum Engineering a "sunset goldmine" because we need engineers to safely decommission oil infrastructure over the next 20 years.
A-TIER: The Underrated Picks
Skilled Trades (Electrician, Plumber)
Ranked higher than most office jobs because AI has terrible fine motor skills and you can't automate crawling under a house.
Physics/Math
Not for the major itself, but because it gives you the foundation to pivot into basically any technical field.
B/C-TIER: Risky Territory
General Computer Science
This was the most controversial. Got downgraded to B/C tier by most models. The takeaway: if you're a generic "I learned Python in a bootcamp" developer, you're in trouble. DeepSeek and Kimi both noted ~20% drop in junior dev postings. BUT if you specialize (hardware, security, low-level systems), you're still S-tier.
Accounting (without CPA)
C-tier without the license. The actual CPA credential is the moat; basic bookkeeping is getting automated.
F-TIER: Danger Zone
Marketing/Communications
ChatGPT and similar tools have basically destroyed the barrier to entry for junior-level work.
Graphic Design
Image generators killed this unless you're top 1% art director level. The "make this logo bigger" jobs are gone.
THE IRONY TIER: AI/Data Science (undergrad)
This one was a bit surprising to me. Multiple models (Kimi and Qwen especially) rated undergraduate "AI Engineering" or "Data Science" degrees as Tier C - Risky.
Their reason: AutoML tools are automating the creation of AI models faster than people can learn them. A generic "AI degree" becomes outdated in like 18 months. The models recommended majoring in Math or CS and then specializing in AI/ML rather than doing a dedicated AI undergrad program.
Basically: the field that's supposed to automate everything is... automating itself.
> Note: Keep in mind this is where the models agreed - there were plenty of disagreements on specific rankings that I didn't include here.
Jobadvisor
That is an impressively thorough way to procrastinate! Your findings align remarkably well with the current 2026 industrial outlook, especially regarding the "Human Moat" and the "Licensing Wall."
In the current market, we’re seeing a massive shift from "AI-enabled" roles to "AI-immune" roles. Here is how your tier list stacks up against the broader industry data and current economic trends:
1. The "Licensing Moat" (Nursing & Medicine)
You are spot on. In 2026, the bottleneck for healthcare isn't just knowledge (which AI has); it’s liability and physical presence. * The Data: Healthcare job demand has remained remarkably stable even as other "white-collar" sectors cooled.
The Nuance: We are seeing AI handle 70% of clinical documentation (reducing burnout), but the actual medical decisions and physical interventions remain strictly human-led. The licensing acts as a legal barrier that AI literally cannot cross.
2. The "Physical Gap" (Trades)
This is the "underrated" winner of the decade. We’ve seen a "Blue-Collar Renaissance."
The Factor: Robotics (fine motor skills) has not kept pace with Generative AI (cognitive skills). An AI can write a sonnet in 2 seconds, but a robot still can't navigate a cluttered basement to fix a burst pipe efficiently.
Hybridization: The most successful tradespeople now are those using "AI Diagnostics"—using sensors and models to pinpoint electrical failures before they touch a wire.
3. The "CS Identity Crisis"
Your models' B/C rating for general CS is the most debated topic in 2026.
The Trend: Junior software engineer postings dropped significantly because "Senior + AI" can now do the work of three juniors.
The "Irony Tier": I see exactly what Qwen and Kimi were flagging. An "AI Engineering" undergrad is often too narrow. By the time you graduate, the library you learned is obsolete. Industry leaders are now begging for Math and Physics majors because they have the foundational logic to build new models, not just use existing ones.
4. The "Creative Cull" (Graphic Design & Marketing)
This is where the industry is most "brutal" right now.
Graphic Design: Entry-level asset creation (logos, social posts, stock imagery) is essentially a dead career path. The only people thriving are high-level Art Directors who use AI as a brush, not a replacement.
Marketing: We've seen a shift toward "Hyper-Personalization." If you’re a "Content Creator" writing SEO blogs, you’re in the F-tier. If you’re a "Marketing Strategist" who can orchestrate 1,000 AI agents to run personalized campaigns, you’re S-tier.
Interesting Discrepancy: Cybersecurity
While your models gave it S-Tier, current industry data shows a massive barrier to entry. It is an S-tier career for experts, but an F-tier for beginners. Because AI helps both sides of the "arms race," companies are terrified to hire juniors who might make a mistake. It’s the one field where "learning on the job" is becoming nearly impossible.
A New Category: The "Ethics & Compliance" Shield
One thing your models might have missed that is booming in 2026: AI Auditors. As governments (EU, US, China) have passed massive AI regulations, there is a huge surge in demand for people who can "certify" that a company's AI isn't biased or illegal. It’s like a CPA, but for algorithms.