◆ Mathematics

What a computational scientist
really does.

20 tasks, each one witnessed by the sources that watched the job — and behind every one, a prompt you can use tonight.

20evidenced tasks
2,030in the US (2025)
$126,710median pay / year
7systems it runs on
See the 20 tasks ↓Open in the atlasThe full career map →
This is what one task looks like here
Perform actuarial computations
Run the actuarial batch for the Q2 portfolio using the validated loss …3 sources agree

The shape of the day

tap a movement to see its tasks

Which one is you, right now?

Pick the moment · no score, no sign-up
Which moment is you right now?
Whichever you pick, the task behind it opens below.

The work, task by task

20 tasks
Hands on the work14
Perform actuarial computations+
Run the actuarial batch for the Q2 portfolio using the validated loss distributions, produce a table of reserve estimates with 95% confidence intervals, flag lines with changes over 10%, and save the numeric outputs and diagnostic logs to the project analysis folder by Friday noon.
jdonetwiki3 agree
when the reply comes backPush once: ask it to sharpen the weakest part, and to say what it assumed. Helpful?
Mentor others on mathematical techniques+
Prepare a one-hour hands-on session for the junior analysts on maximum likelihood and bootstrapping: draft the teaching notes, build two short worked examples with datasets, and upload the slides and notebooks to the team folder before Wednesday so they can prep.
jdonetwiki3 agree
when the reply comes backPush once: ask it to sharpen the weakest part, and to say what it assumed. Helpful?
Design mathematical models to interpret data+
Develop a parsimonious model linking sensor covariates to outcome rates: specify the equations, fit competing variants to last year’s data, compare AIC and predictive error, and export the chosen model specification and parameter table to the shared models repository.
jdwiki2 agree
when the reply comes backPush once: ask it to sharpen the weakest part, and to say what it assumed. Helpful?
Assemble sets of assumptions, and explore the consequences of each set.+
Create five coherent assumption sets for the coastal circulation model varying boundary forcing and turbulence closure, run each set through the model, compare time-averaged transport and error metrics, and produce a short memo recommending which assumption set to use for the Phase II proposal.
onet
when the reply comes backPush once: ask it to sharpen the weakest part, and to say what it assumed. Helpful?
km/h RPMAddress the relationships of quantities, magnitudes, and forms through the use of numbers and symbols.+
Derive and typeset the governing equations linking scalar variance to dissipation rate for our mixing study, solve the linearised case symbolically, produce numeric examples with parameter values from table 2, and export the derivation and plots for inclusion in the methods section.
onet
when the reply comes backPush once: ask it to sharpen the weakest part, and to say what it assumed. Helpful?
Design, analyze, and decipher encryption systems designed to transmit military, political, financial, or law-enforcement-related information in code.+
Design and test a prototype cipher framework for secure telemetry between sensor nodes and the control server, document threat model and key management, run the test vectors and report performance overhead and failure cases for the classified appendix.
onet
when the reply comes backPush once: ask it to sharpen the weakest part, and to say what it assumed. Helpful?
Compile charts and graphs for data presentation+
Produce publication-quality charts showing ensemble mean, standard deviation, and outlier trajectories for the June run, use consistent color palette and axis ranges, export PNG and vector versions, and attach a short legend-ready caption for each figure.
jd
when the reply comes backPush once: ask it to sharpen the weakest part, and to say what it assumed. Helpful?
Identify problems through discussion with colleagues or clients+
Schedule and lead a one-hour problem-discovery meeting with Sam in modelling, Aisha in instrumentation, and Marco in operations, collect specific failure examples, list hypotheses, assign two follow-ups with deadlines, and circulate minutes within 24 hours.
jd
when the reply comes backPush once: ask it to sharpen the weakest part, and to say what it assumed. Helpful?
Grow the practice4
Watch and assess2

What the work runs on

named inside the evidenced tasks
9 tasksC++implements fast simulation code to test new mathematical estimators across many draws
5 tasksLinuxruns numerical code and batch jobs for heavy actuarial computations and manages output files
4 tasksIBM SPSS Statisticscreates worked examples, runs demonstrations of estimation and bootstrap procedures for teaching
2 tasksJavaScriptorchestrates data transformations and quick model fitting in interactive notebooks and visual comparisons of fit
2 tasksBashorchestrates symbolic solver calls, runs numeric experiments, and assembles output files for export
1 taskC#implements prototype encryption routines and test harnesses for performance and correctness checks
1 taskAtlassian JIRAcaptures issues, assigns follow-ups, and tracks deadlines raised during the meeting
Browse every evidenced software task →

The same task, four heights

this page is height one
ExecuteDo today's task, with fewer mistakesyou are here → ImproveMake it easy for the next person to acceptin the atlas → DecideWork out the right move when it is unclearin the atlas → BecomeLearn the pattern so it stops coming backin the atlas →

Can AI actually do this job?

the honest answer

It can

where it genuinely helps
  • Explain the theory behind the work
  • Draft, tidy and structure your writing
  • Rehearse a hard conversation before you have it
  • Build a study plan that fits your gaps

It cannot

where it stops, completely
  • Be in the room where a computational scientist actually works
  • Carry the responsibility when the call is wrong — that weight stays yours
  • Notice what no one wrote down: the hesitation, the thing left unsaid
  • Live with the outcome
M · R6 of R7This job sits on calculating — real theory, real diagnosis, but reality still holds the grading pen.

What the work pays

two countries, two different measures

United States

this exact occupation · BLS 2025
  • $126,710 a year — the middle: half earn more, half earn less
  • The lowest tenth earn near $69,240; the top tenth near $195,190
  • 2,030 people employed in this occupation

India

the occupation GROUP, not this job · PLFS via ILOSTAT 2025
  • ₹38,298 a month — the median for Professionals, the group this work sits in
  • India publishes pay by broad occupation group, so this covers many jobs besides this one. It is a shape, not a salary.
read this carefullyThese two numbers are not comparable and must not be converted into each other. One is a yearly figure for this job alone; the other is a monthly figure for a whole family of jobs. What travels between them is the pattern, not the amount: experience lifts pay almost everywhere.

Where the evidence lives

open any of it yourself
O*NETThe US Department of Labor's own record of this job7 of these tasks ESCOThe European skills and occupations register4 of these tasks Job descriptionsReal postings, in the wild8 of these tasks WikipediaHow the world describes this work8 of these tasks US Bureau of Labor StatisticsPay and employment, by occupation2025 figures

Close to this work

12 nearby
MathematicsSimulation Engineer25 evidenced tasks MathematicsOperations Research Analyst23 evidenced tasks MathematicsGame Theorist23 evidenced tasks MathematicsMath Content Writer22 evidenced tasks MathematicsMathematician20 evidenced tasks MathematicsApplied Mathematician20 evidenced tasks MathematicsTheoretical Mathematician20 evidenced tasks MathematicsQuantitative Analyst Quant20 evidenced tasks

Questions people actually ask

You usually split time between coding models, running simulations, and reading results. Morning might be debugging a C++ or Python model, then running it on Linux servers with Bash scripts to manage jobs.

Afternoons often mean analyzing output with tools like IBM SPSS Statistics or custom scripts, plotting charts and graphs, and meeting colleagues to identify problems and set the next experiments.

Start with C++ for heavy numerical code, plus Bash for running jobs on Linux. JavaScript or C# helps if you build visualizations or GUIs, and JIRA is used for task tracking.

Also learn IBM SPSS Statistics or equivalent for some statistical analyses; many teams mix custom code with off‑the‑shelf stats tools.

According to the U.S. Bureau of Labor Statistics (BLS), about 2,030 people were employed in this role in 2025. The median wage was $126,710 per year.

The lowest 10% earned about $69,240, and the top 10% earned about $195,190, per BLS data. Your pay depends on sector, experience, and whether you’re in industry, government, or academia.

Use AI models as tools for pattern finding, code suggestions, or preliminary data cleaning, but never as final proof. Always verify outputs by running reproducible experiments and cross‑checking with known methods.

Keep sensitive data off public models, document model versions and parameters, and treat AI suggestions like hypotheses you must test with your experiments and statistical analysis.

Focus on core math (calculus, linear algebra, probability), programming (C++ plus Bash on Linux), and basic statistics (use IBM SPSS Statistics or free alternatives). Build small projects: design a model, collect sample data, and analyze it.

Practice by designing experiments, assembling assumptions, and exploring consequences—write short reports and learn to present results. Mentoring or study groups help accelerate learning.

Computational scientists build and test mathematical models for physical or engineering problems, often writing high‑performance C++ code and running experiments on Linux. They focus on applying math to scientific questions.

Data scientists often focus more on business data, productionML, and pipelines; mathematicians emphasize proofs and theory. Computational scientists sit between: they prove ideas, code models, and run experiments.

Model design and experiment design are key: assemble assumptions, design experiments, and interpret consequences. You must collect data carefully and make plots to convey findings.

Analytical thinking, deriving corollaries from theorems, and clear writing for reports or conference presentations matter every week. Mentoring juniors and discussing problems with colleagues are also regular tasks.

km/h RPM

How far can AI take you?

one question
Be honest — how far can a language model take you on calculating?
There is a real answer, and it is not the flattering one.