$70-$180/hr data science and machine learning work, on your schedule
Review AI notebooks and model evals for the data leakage, the p-hacking, the silently wrong baseline. The instinct for what makes a result trustworthy. Paid hourly, remote, a few hours a week.
Trusted by top research companies
Hi, we're Zac and Jack, the founders of Terac. We want to talk to you directly, because you are the most important part of what we're building.
Terac is a community of experts. People who have spent years getting good at something specific and hard. The world is about to need more of you, not less. As AI takes on more of the world's work, the bottleneck shifts to the people who actually know what they're talking about.
Expert labor is the rarest resource in the world right now, and it is shockingly hard to find. The companies that need a data scientist's eye on a leaky validation set spend weeks chasing people, paying placement fees, and settling for whoever is available. Meanwhile thousands of qualified people are sitting with knowledge that no one ever asks for.
That gap is what we're here to close. Every project that lands on Terac is routed to the people who actually know the answer, on their schedule, paid fairly, and only when the work is verified. No middleman taking a cut of your time. No vague gigs. No chasing checks.
We care about every single person in this community. If you join Terac, you're not a row in a database to us. We read the feedback. We answer the emails. We will fight for you when a customer is being unreasonable, and we will be honest with you when something on our side is broken. The quality of this panel is our entire company, and we owe you a serious bar.
If you've made it this far, here is what we're asking: claim your profile. Put your expertise on the record. Let the world's most ambitious teams come find you for the work only you can do.
Data Science questions
Still curious? Write to us at support@terac.com.
Narrow sub-specialties are among the most in-demand profiles. Models trained on generic data science content consistently fail at field-specific reasoning, so experts who can evaluate causal identification, embedding fine-tuning, or forecasting pipelines beat generalists. Deep experience in one sub-domain is an asset, not a limitation.
It varies, but typically AI Python or R analyses, model evaluation write-ups, feature engineering rationale, and statistical methodology explanations. You may also write worked examples showing how you would approach a problem, like choosing a regularization strategy or diagnosing data leakage, so the model learns your reasoning.
Credentials like the CAP or an advanced degree in statistics, applied math, or a quantitative social science help route you to projects needing verified authority, especially regulated finance or healthcare work held to a higher rigor bar. They are not required, but listing them gets you considered for higher-complexity assignments.
No. All data is fully synthetic, publicly licensed, or anonymized before it reaches you, and no task requires processing, storing, or transmitting real personal data. Briefs specify the data context upfront, so any healthcare or financial scenario is clearly simulated, and your participation never conflicts with data protection rules you are bound by.
Yes, and it is genuinely underrepresented. Much AI training content skews academic or Kaggle-style, so reviewers with SQL-heavy pipelines, distributed compute, or modern data stack experience provide coverage that is otherwise hard to source. Tasks on data-engineering-adjacent reasoning, scalability tradeoffs, or enterprise ML deployment are a strong match.
Why your expertise matters
A model's analysis is statistically plausible and analytically wrong: data leakage between splits, a bad feature engineering choice, a misapplied hypothesis test a non-specialist accepts at face value. Only a working practitioner who has shipped production models and defended methodology can tell rigor from confident-sounding hallucination. Your reviews teach the model that line.
How pay works
Pay in the $70-$180 band tracks your sub-specialty: production experience in causal inference, time-series forecasting, or ML model risk lands at the upper end. Work is remote, billed by verified hour or completed task, and paid only after your submission meets the quality bar. No retainers or minimums.
What the work looks like
A sample of the data science and machine learning work you would pick up. Every project is scoped, remote, and paid on verified completion.
- Review a model's churn-prediction notebook and flag train/test data leakage, bad imputation, or metric choices that would mislead a product team.
- Evaluate an AI explanation of p-values and confidence intervals for a non-technical audience, correcting any overstated certainty or conflated significance.
- Write a worked example of a proper difference-in-differences analysis on a synthetic dataset, annotating each assumption and diagnostic check.
- Assess a model's feature importance report and check whether it distinguishes permutation importance from SHAP and whether the numbers support the claims.
- Stress-test a machine-produced time-series forecast with adversarial inputs like regime changes or sparse seasonal data, documenting where it breaks.
- Judge whether an AI SQL cohort retention query handles edge cases like users with multiple first-event timestamps or sessions spanning UTC midnight.
Specialties we match
Data Science projects span a wide range of focus areas. Tell us where you go deep and we route the work that fits.
- Feature engineering and selection
- Experiment design and A/B testing
- Causal inference (DiD, IV, RDD)
- Time-series forecasting (ARIMA, Prophet, TFT)
- ML model validation and diagnostics
- SQL and dbt pipeline authoring
- Scikit-learn, XGBoost, LightGBM
- Deep learning with PyTorch or TensorFlow
- Model monitoring and drift detection
- Statistical hypothesis testing
- Survival analysis
- NLP and text classification
