Data Science Expert, Python, SQL And GenAI

The work

You will author and verify original data science problems based on real analytical work across industries such as telecom, finance, government, e-commerce, and healthcare. Problems should include realistic business goals such as fraud detection, forecasting, optimization, risk analysis, and customer analytics.

You will create end-to-end Python problems covering data ingestion, cleaning, exploratory analysis, feature engineering, modeling, validation, and deployment considerations. You will also write clear prompts, answers, summaries, documentation, and evaluation criteria, then verify solutions with standard data science tools.

• Write computational problems that cannot reasonably be solved by hand.
• Use deterministic methods, including fixed random seeds when needed, so solutions are reproducible.
• Evaluate responses and code for correctness, clarity, and business relevance.
• Work with text-based tasks involving question answering, text generation, summarization, coding, and response evaluation.
• Consider scalable processing, partitioning, performance, and memory limits in problem design.

What it pays and takes

This is project-based, part-time contractor work rather than a permanent position. The role is intended for experienced data scientists who can explain technical work clearly in written English.

• Pay: $15 to $40 per hour, paid in USD.
• Location: Worldwide.
• Language: Fluent English, with written proficiency at C1 level or higher.
• Experience: At least five years of hands-on data science work with measurable business impact.
• Availability: About 10 to 20 hours per week during active project phases; the structured role details indicate a 20-plus-hour weekly requirement.
• Expert Python skills, including Pandas, NumPy, SciPy, scikit-learn, and statsmodels.
• Expert SQL skills, including complex joins, aggregations, window functions, and database operations.
• Strong statistics and machine learning knowledge, including feature engineering, model selection, evaluation, and error analysis.
• Experience with visualization tools such as Matplotlib; Seaborn is a plus.
• Experience designing workflows from ingestion through cleaning, analysis, modeling, and validation.
• Experience with LLMs, retrieval-augmented generation, prompt engineering, and vector databases.
• Understanding of MLOps and model deployment, including packaging, reproducibility, and basic monitoring.
• Familiarity with TensorFlow or PyTorch; LangChain experience is a bonus.

How it works

Apply on OpenTrain. The employer reviews applications there.

About AI training work

AI training work uses human-written examples and reviews to improve how artificial intelligence systems respond, reason, and complete tasks. Experienced specialists are needed to create reliable examples, check difficult answers, and bring real subject knowledge to model development.