AI Trainer (Earth & Environmental Sciences)
micro1 • Remote
Education
Not stated
Type
Hourly
Pay
$60-120/hr
Your rate depends on your experience and the task.
Actively hiring
100 openings
Checked
Only apply if this role fits your skills and interests.
About this role
From the micro1 listing
Pay: $60–$120/hour equivalent, paid per accepted task Location: Global, fully remote Job Type: Contractor Time Commitment: 15+ hours per week Schedule: Flexible micro1 is looking for Computational Earth and Environmental Scientists with prior professional AI-training or model-evaluation experience to support a project focused on improving advanced AI systems. You will use your scientific expertise and programming skills to create challenging technical tasks, evaluate model-generated code and analyses, identify failures, and establish objective standards for correct solutions. This role is intended for candidates who have completed substantive AI-training work involving earth, climate, environmental, or geospatial content.
What You’ll Work On
- Create realistic computational earth-science tasks that challenge advanced AI models.
- Develop task instructions, reproducible environments, reference solutions, and evaluation criteria.
- Evaluate model-generated code, geospatial analyses, simulations, forecasts, and technical deliverables.
- Identify issues involving coordinate systems, spatial or temporal data, model assumptions, uncertainty, numerical methods, or reproducibility.
Required Qualifications
- Prior paid experience training, evaluating, benchmarking, or improving AI systems on scientific or technical tasks.
- Strong practical proficiency with Python.
- Deep expertise in at least one relevant earth or environmental-science specialization.
- Hands-on experience using computational methods for geospatial analysis, modeling, simulation, forecasting, remote sensing, statistics, or reproducible research. Relevant specializations may include geoscience, geology, geophysics, atmosphere and climate, meteorology, ocean and marine science, hydrology, remote sensing, environmental science, and sustainability.
Public Proof of Skill
Strong applicants should provide evidence such as:
- Peer-reviewed publications or preprints involving computational methods
- Public geospatial, climate, ocean, or environmental analysis code
- Google Scholar, ORCID, or equivalent research profiles
- Open-source scientific software contributions
- Public datasets, models, notebooks, maps, or simulations
- Publicly documented forecasting, remote-sensing, or environmental projects
Preferred Qualifications
- Experience creating AI-training tasks, reference answers, rubrics, graders, or benchmark problems.
- Experience reviewing other AI trainers’ technical submissions.
- Experience with NumPy, SciPy, pandas, xarray, GeoPandas, GDAL, rasterio, NetCDF, or comparable tools.
- Experience validating computational results against observations, measurements, or established models.
- Availability to begin immediately.
Application Process
- Submit your application, AI-training experience, specialization, and public technical evidence.
- Complete an AI interview containing technical and integrity checks.
- Complete a hiring-manager review.
- If selected, complete project onboarding and begin your first assigned task.
Compensation
- Compensation is output-based. Experts are paid for tasks that meet project specifications and quality requirements. Task availability and completion time may vary.
Requirements
- Earth & Environmental Sciences
- AI Training
- Programming Skills
How long hiring takes
On the platforms we refer people to, half of hires came within about 25 days of applying. Time from applying to hire:
AI Trainer interview prep
Questions AI Trainer interviewers often ask. They aren't taken from this listing.
How do you evaluate whether a model's response to a prompt satisfies the instruction, versus just sounding plausible?
I go back to the instruction line by line and check each requirement against the response individually, rather than reading the response holistically and asking if it feels right. A confident, well-written response can still quietly skip a constraint the prompt asked for.
What's your process for writing a rating rubric that two different reviewers would apply the same way?
I define each rating level with a concrete example rather than an abstract description, since two reviewers can read the same adjective differently but usually agree on whether a response matches a given example. I test the rubric on a handful of edge cases with another reviewer before rolling it out.
How do you decide when a prompt itself is ambiguous versus when the model's response is simply wrong?
I ask whether a reasonable person could interpret the prompt more than one way and still be following the instruction correctly. If the response fails under every reasonable reading, it's a model error; if it only fails under one specific reading, the prompt needs to be flagged as ambiguous instead.
How do you document the reasoning behind a rating so someone reviewing your work later understands the decision?
I write a short note pointing to the specific part of the response that drove the rating, so a reviewer can see what I was reacting to. A rating with no reasoning attached is hard to audit and even harder to disagree with productively.
Why this role
This AI Trainer (Earth & Environmental Sciences) role pays $60-120/hr for the STEM background you've already built.
Talent pool
We're light on STEM candidates
We've tracked 1,036 STEM listings and have few candidates to put forward for them. Set up a profile and we'll consider you for roles like this one.
Set up your profileSkills and categories
More jobs in these fields:
Related jobs
Verified reviews
Community reviews
Leave your review
Common questions
What does the micro1 application process look like?
You start with a video interview with Zara, micro1's AI interviewer. Treat it like a video call: good light, clear sound and spoken answers to technical questions. A person reviews the recording, and some roles add a short skills test.
Can I apply to micro1 from outside the US or UK?
Yes. micro1 accepts applicants from most countries. Most roles need strong written and spoken English.
Is research-level AI training the same as data labeling?
It's closer to academic research. You write or check proofs, solve advanced problems and test the logic behind a model's step-by-step reasoning in your field.
Do I need a PhD for research-level AI training roles?
The best-paid roles usually expect a PhD or current enrollment. The domain assessment decides, though, and if you can solve the problems the degree matters less.
What does STEM work look like for an AI Trainer (Earth & Environmental Sciences)?
You check AI answers to STEM questions, correct their errors and write expert reasoning. As an AI Trainer (Earth & Environmental Sciences), you're hired for that judgment, so hands-on STEM experience matters.
What specific skills does this listing call for?
Earth & Environmental Sciences, AI Training, and Programming Skills are named directly in the listing. If you don't have hands-on experience with these, expect the screening to test for them.
What does the 'openings' count on this listing mean?
It's how many people micro1 still wants to hire for this role: 100 when we last checked. A high number means more places to fill. The screening stays the same.
How much does this specific role pay?
This listing is posted at $60-120/hr, an hourly rate. Where you land in the range depends on your background and the assessment. Pay can change after we check the listing, so confirm the rate on the platform's application page.
What happens when I click Apply on this listing?
Apply opens micro1's site, where you create an account and follow its sign-up steps.