The Frontier AI Data Scientist Challenge

Design, construct, and execute a comprehensive evaluation of a state-of-the-art AI agent performing a real-world data science workflow.

Submission Deadline: April 10, 2026 at 11:59 PM UTC

1. Background & Objective

Frontier AI models have demonstrated remarkable capabilities in coding, logic, and reasoning. But how effectively can they operate as human data scientists?

This competition challenges participants to design, build, and fully execute a comprehensive evaluation of one of the frontier AI Agents (Claude, Gemini, ChatGPT, Grok) performing a data science workflow. This is not a proposal or planning exercise. You must deliver a complete, end-to-end evaluation.

No rigid rubric or predefined dataset is provided. The goal is to define what a rigorous, real-world AI evaluation looks like and execute it.

2. The Challenge

Participants are expected to approach AI evaluation thoughtfully and creatively. The domain and methodology are fully open-ended, but submissions must follow these principles:

  1. End-to-End Execution: Fully design and execute the evaluation with results and findings.
  2. Realistic Task Construction: Tasks should reflect real-world data science work.
  3. Real Data: Use real datasets, not synthetic or toy data.
  4. Realistic Evaluation: Metrics and assessments should mirror real-world evaluation standards.
  5. Comprehensiveness & Generalizability: Avoid narrow tasks; results should generalize to real-world scenarios.

Note: The challenge is intentionally open-ended to encourage creativity.

3. What We Value (Judging Criteria)

Submissions will be evaluated based on depth and execution:

  1. Creativity & Task Design: Innovation and realism of the challenge.
  2. Quality of Evaluation: Rigor in measuring performance and limitations.
  3. Organization: Clarity, structure, and reproducibility of the submission.

4. Submission Requirements

All submissions must include everything below via one .zip file:

  1. Final Report (PDF): Full description of methodology and findings.
  2. Appendices & Assets: Datasets, prompts, specs, scripts, and all supporting materials.
  3. Standard Formats: Files should be in formats like .csv, .json, .md, .txt.
  4. Structured Directory: Clean, logical organization is required.
  5. File Size Limit: Maximum 100 MB, submitted as a single .zip.

5. Awards

Top 3 participants receive:

Winners will be notified by April 30, 2026.

Submit Your Entry

6. Terms & Conditions

6.1 - Eligibility

  1. You must be at least 18 years old and/or the age of majority in your jurisdiction to participate.
  2. Participation is void where prohibited by law. No purchase necessary. This challenge is not open to residents of Florida, New York, or Quebec.
  3. Employees, contractors, or affiliates of Turing Enterprises, Inc. ("Turing") are not eligible to win rewards.
  4. Turing reserves the right to verify eligibility before granting any prize or reward.
  5. Individual Participation: Each submission must represent the work of a single individual. Team entries are not accepted. Each individual may participate in only one (1) entry. If Turing determines that an individual appears in more than one entry, Turing may disqualify any or all affected entries in its sole discretion.

6.2 - Submission Requirements

To qualify for a successful submission, the entry must:

  1. Be in English.
  2. Be your original work created by you and not infringe or violate the rights of any person or entity, including their intellectual property, publicity, or privacy rights.
  3. All submissions must include everything below via one .zip file:
    • Final Report (PDF): Full description of methodology and findings.
    • Appendices & Assets: Datasets, prompts, specs, scripts, and all supporting materials.
    • Standard Formats: Files should be in formats like .csv, .json, .md, .txt.
    • Structured Directory: Clean, logical organization is required.
    • File Size Limit: Maximum 100 MB, submitted as a single .zip.
  4. Utilize real data (not synthetic or toy data).
  5. Participants represent and warrant they have the legal right to use and submit such data and that it does not contain personally identifiable information (PII).
    • Submissions should not include confidential, proprietary, or non-public information unless you have express written authorization to disclose it.
  6. Reflect a realistic, end-to-end execution of an AI agent evaluation, demonstrating rigor, reproducibility, and generalizability.
  7. Not have been autonomously generated through the use of robotic, repetitive, automatic, or programmed entry methods.

6.3 - Review Process

  1. Submissions will be reviewed via a panel of experts who will evaluate entries based on:
    • Creativity & Task Design
    • Quality of Evaluation, and
    • Organization.
  2. If your submission is selected as a winner (1st, 2nd, or 3rd place), you may be eligible to receive a monetary reward as specified above.
  3. Verification Interviews: Judges reserve the right to request a virtual interview with any participant to discuss their submission. Participants must be prepared to explain their methodology, decisions, and findings in detail. Inability to demonstrate thorough understanding of your own submission negatively impacts your submission evaluation.
  4. Turing retains final and sole discretion on what is considered a successful submission and decisions are final and binding on all matters.
  5. Turing reserves the right to deny or revoke rewards in cases of fraud, misrepresentations of your age/identity, ineligibility, duplicate submissions, plagiarism, or violation of these terms.
  6. Winners will be notified via email to the individual who sent in the submission.

6.4 - Payment

  1. If your submission is selected as a winner, any prize payment will be made by Turing or its designated third-party payment provider. Turing may contact you by email to collect the information and documentation required to process payment.
  2. To receive payment, winners must timely provide all information reasonably requested by Turing or its payment provider, which may include valid banking information and applicable tax documentation, including a completed Form W-9 for U.S. persons or Form W-8BEN for non-U.S. persons. For U.S. winners, Turing may also issue an IRS Form 1099, as required by applicable law.
  3. Payments will generally be made within ninety (90) days after winners are announced, subject to timely completion of all payment onboarding, verification, and documentation requirements.
  4. Turing may request additional information to verify your identity and eligibility before disbursing any rewards.
  5. All federal, state, and local taxes, fees, and surcharges associated with any prize or reward are the sole responsibility of the participant. Failure to provide requested tax documentation may result in forfeiture of the prize.
  6. Rewards are non-transferable and no substitution of an entrant or cash equivalents is permitted.

6.5 - Limitations of Liability

  1. Rewards are offered "as is" with no warranties.
  2. Turing may modify, suspend, or cancel the program at any time without prior notice.
  3. To the fullest extent permitted by law, Turing and its respective affiliates, officers, directors, employees, and agents (collectively, the "Released Parties") shall not be liable for any loss, claim, damage, liability, cost, or expense of any kind arising out of or in connection with participation in the program.
  4. In no event shall the total liability of the Released Parties to any participant exceed one hundred U.S. dollars ($100 USD), regardless of the cause of action.
  5. Participants expressly waive and release any and all claims against the Released Parties arising out of or relating to the program.
  6. Participants hereby agree to indemnify Turing against any and all claims from any third party for breach of these terms and conditions or any use by Turing of a submission.

6.6 General/Miscellaneous

  1. You grant Turing a worldwide, royalty-free, perpetual, irrevocable, sublicensable, and transferable license to use, reproduce, publish, modify, and publicly display your submission (including reports, scripts, and data) in connection with artificial intelligence (AI) research, development, marketing, or product development, including the right to share the submission materials with Turing's affiliates, clients, and partners.
  2. Participants are solely responsible for identifying, understanding, and complying with all terms, conditions, licenses, usage restrictions, and other obligations applicable to any third-party materials or services used in connection with a submission, including any datasets, APIs, open-source software, model providers, benchmark licenses, cloud tools, libraries, frameworks, or other third-party content, software, or services. Each participant represents and warrants that its submission, and Sponsor's receipt, review, and use of the submission as contemplated by these terms and conditions, will not violate any applicable third-party terms, licenses, contractual restrictions, or legal requirements.
  3. Turing collects and uses personal data in accordance with its Privacy Policy.
  4. Except where prohibited, participation in the challenge constitutes your consent to Turing's use of your name, likeness, and entry for promotional purposes in any media, worldwide, without further payment or consideration.
  5. These terms and any disputes arising out of or relating to this program shall be governed by and construed in accordance with the laws of the State of California, without regard to its conflict of law principles.
  6. You agree that any legal action shall be brought exclusively in the state or federal courts located in Santa Clara County, California, and you hereby consent to the personal jurisdiction and venue of such courts.

By submitting, you agree to the terms and conditions described above.