> Markdown version of [/jobs/ext/4704474-remote-ai-trainer-data-science](https://www.wearedevelopers.com/jobs/ext/4704474-remote-ai-trainer-data-science). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Remote AI Trainer, Data Science

- **Company:** ACCOUNTINGFLY
- **Location:** Philadelphia, PA, United States (Remote available)
- **Experience:** Experienced
- **Salary:** $83,200 - $260,000
- **Contract:** Permanent contract
- **Skills:** Artificial Intelligence, Data Analysis, Computer Vision, Information Engineering, Information Leak Prevention, Python (Programming Language), Machine Learning, Raw Data, Standard Sql, SQL Databases, Large Language Models, Model Validation
- **Published:** November 13, 2026
- **Apply:** https://www.thejobnetwork.com/job/76186976-44e0-4955-8767-f407bc5d645f/remote-ai-trainer-data-science

## About the Role

* 2+ years of applied data science experience preferred (analytics, ML, statistical modeling, experimentation, or analysis-adjacent data engineering). Undergraduate study doesn't count, but strong practical work can offset less experience.
 * Current or recent hands-on experience as an individual contributor
 * A bachelor's degree in a quantitative field, completed or in progress. An advanced degree is a plus but not required.
 * Working proficiency in Python and/or SQL
 * Depth in at least one area: business analytics/BI, applied ML, statistics and experimentation, forecasting, data engineering, or applied NLP/computer vision
 * Experience owning a multi-step analysis end-to-end, from raw data to a decision
 * The ability to judge whether results are valid (leakage, confounding, selection effects, multiple testing, power) and explain them in business or scientific terms
 * Familiarity with AI/LLM tools and a sharp eye for answers that sound right but aren't
 * Full professional or native-level English with strong written communication
 * Strong attention to detail, comfort with ambiguity, and the ability to work independently from written instructions

## Description

We're looking for experienced data scientists to test how well advanced AI models handle real data science work. That means messy-data analysis, experiment design, model validation, and SQL/Python development. You'll design realistic tasks from your own experience and grade the AI's output against the standard you'd hold a colleague to.

The work is fully remote and self-directed. Clear written reasoning matters as much as technical depth.

What You'll Do

 * Design realistic tasks drawn from your own work, such as exploratory analyses, A/B test readouts, model validation, SQL pipelines, and forecasts. Each task includes prompts and supporting files like datasets, schemas, notebooks, and stakeholder briefs.
 * Put frontier AI models to the test by running your tasks and evaluating the analyses, code, models, and write-ups they produce against a professional standard.
 * Compare outputs head to head, decide which model performed better, and document where each one fell short.
 * Write detailed grading rubrics and tests that define what a correct deliverable must include, with written pass/fail justifications.
 * Call out concrete failures with evidence, including data leakage, confounding, incorrect joins, misapplied tests, and off-brief interpretations.
 * Review and refine tasks built by other experts., 3. Onboard and train. Pass the assessment, then complete onboarding and a short paid training project.
 4. Propose a task from your own experience and get expert feedback.
 5. Build and test. Once your proposal is accepted, build the full task and test it against frontier models.