Competitive Data Scientist

Competitive Data Scientist

Employment Type: Contractor
Work Arrangement: Fully Remote
Eligible Locations: Global
Expected Commitment: Approximately 15 hours per week
Schedule: Flexible

Compensation

Hourly Rate: $60–$120 per hour

Annualized Part-Time Equivalent: $46,800–$93,600 per year

The annualized figure assumes 15 hours per week and 52 weeks per year. It is for comparison only; actual earnings depend on project availability and hours worked.

About the Role

We are seeking experienced Competitive Data Scientists with backgrounds in data science, machine learning, applied mathematics, or quantitative modeling, particularly those with experience in data science competitions such as Kaggle or DrivenData.

You will design challenging technical tasks that require sound mathematical reasoning, practical coding ability, and verifiable solutions. Your work will involve preparing reproducible environments, implementing reference solutions, developing reliable evaluation tests, and investigating how coding agents approach and solve complex problems.

This opportunity is suited to professionals who can go beyond producing plausible answers to demonstrate correctness through rigorous implementation, testing, and validation.

Key Responsibilities

Task Development & Environment Setup

• Design challenging data science and machine learning tasks based on practical technical problems.

• Build reproducible environments with the required data, dependencies, and configuration.

• Implement reference solutions that demonstrate correct approaches to assigned problems.

• Document task requirements, assumptions, and expected outcomes clearly.

Testing & Evaluation

• Develop tests and evaluation criteria that distinguish correct solutions from incomplete or flawed approaches.

• Verify reference solutions and ensure results can be reproduced.

• Establish reliable checks for assessing task completion and solution correctness.

• Identify weaknesses in evaluation methods and improve them where necessary.

Coding Agent Assessment

• Run coding agents against assigned tasks and investigate their performance.

• Analyze failed attempts to understand errors in implementation, reasoning, or assumptions.

• Refine tasks based on observed agent behavior.

• Identify data leakage, incorrect assumptions, implementation errors, and other factors that affect solution validity.

Technical Documentation & Quality Assurance

• Explain solution reasoning and verification procedures clearly.

• Document relevant assumptions, implementation details, and testing steps.

• Review tasks for technical accuracy, reproducibility, and consistency.

• Apply rigorous problem-solving and validation methods throughout the task development process.

Required Qualifications

• Demonstrated experience in data science, machine learning, quantitative modeling, applied mathematics, or relevant research competitions.

• Experience with competitive data science platforms such as Kaggle, DrivenData, or similar environments.

• Strong mathematical reasoning and analytical problem-solving skills.

• Practical programming ability sufficient to implement, test, and verify solutions independently.

• Ability to evaluate technical approaches and identify flawed assumptions or incorrect results.

• Strong understanding of the methods and techniques used in your area of expertise.

Preferred Qualifications

• Strong programming experience, particularly in Python.

• Experience with software testing, debugging, and unfamiliar codebases.

• Participation in data science or machine learning competitions through platforms such as Zindi, AIcrowd, Numerai, DataSource.ai, or CodaLab/Codabench.

• Experience developing reproducible computational environments and reference implementations.

• Familiarity with evaluating coding agents and diagnosing failures in generated code.

Key Skills

• Competitive Data Science

• Machine Learning Engineering

• Data Science

• Applied Mathematics

• Quantitative Modeling

• Python Programming

• Coding and Debugging

• Reference Solution Development

• Reproducible Environments

• Testing and Validation

• Evaluation Criteria Design

• Coding Agent Evaluation

• Data Leakage Detection

• Statistical Reasoning

• Algorithmic Problem-Solving

• Technical Documentation

• Quality Assurance

Work Environment

This is a fully remote, hourly contract opportunity with an expected commitment of approximately 15 hours per week.

The schedule is flexible, allowing you to choose your working hours and days, including weekends if preferred. Work focuses on developing technical tasks, implementing and verifying solutions, and evaluating coding agent performance.

What Success Looks Like

Success means creating challenging, reproducible data science tasks with verified reference solutions and reliable evaluation criteria.

Your work should make it possible to distinguish genuinely correct solutions from plausible but flawed approaches, while identifying implementation errors, data leakage, and weaknesses in coding agent performance.