MLE Bench: Data Analyst

What you'll do

Where the engineer version of this project builds and fixes ML systems, the analyst version interrogates what they produce. You analyze structured and unstructured data from ML training, inference and evaluation pipelines, define and validate the metrics used to judge model performance, and look into data distributions, outputs, failure modes and edge cases. Python and SQL are the tools, for analysis, reports and evaluation workflows. You also check data quality and consistency across datasets and experiments, produce documented, reproducible analyses, and work with ML engineers and researchers to design demanding evaluation scenarios for MLE Bench.

Who fits

• 3 or more years as a data analyst or analytics-focused engineer
• Strong Python for analysis and solid SQL on relational data
• Experience analyzing ML outputs and evaluation metrics
• A good grounding in statistics and analytical reasoning
• Comfort with large, complex datasets and drawing reliable conclusions from them
• Clean, well-documented analytical code and excellent spoken and written English

What it pays

No rate is published in the ad. Turing sets rates per project; see Turing.

Worth knowing

Good:

• A route into AI evaluation for analysts that does not require deep-learning engineering
• Terms are stated: at least 4 hours a day, minimum 20 hours a week, 4 hours of PST overlap, 3 months adjustable based on engagement
• Vetting is a 60-minute technical interview with a live coding challenge

Less good:

• Restricted to India, Pakistan, Nigeria, Kenya, Egypt, Ghana, Bangladesh, Turkey, Brazil and Mexico
• Contractor assignment, no medical or paid leave
• No pay figure
• The 20-hour minimum and PST overlap limit who can fit it around other work

The engineer version of the project is listed separately.

About this listing

Read from Turing's job board on 18 October 2026; first posted January 2026. Requirements and terms are Turing's; the contrast with the engineer role and the caveats are ours. See Turing.