Coding Expert - SQL

Coding evaluation work

Help improve AI coding systems through practical SQL expertise. The role combines evaluating generated code, writing solutions, and explaining the reasoning behind technical decisions.

Assignments

• Check generated queries for correct results, performance, and sound database practice.
• Write queries for different relational database systems.
• Explain the approach and reasoning in accessible English.
• Identify mistakes in joins, filters, aggregation, and assumptions about schemas.

Qualifications

Applicants need fluent English, clear analytical thinking, and the ability to explain technical concepts in writing. Required technical knowledge includes complex queries, indexing, and database normalization.

A bachelor’s degree in Computer Science, Data Science, or a related field is required or may be in progress.

Three to five years working with MySQL, PostgreSQL, SQL Server, or Oracle is helpful rather than mandatory. Experience in data labeling, reinforcement learning from human feedback, or other AI training work is an additional advantage.

Schedule and compensation

The engagement is an hourly contract on an at will basis. Work is fully remote and asynchronous, with at least 15 hours per week required; within that commitment, you organize your own schedule. Contributors based in the United States are preferred.

The published US pay range is $25 to $100 per hour. Starts are arranged on a rolling basis as projects launch and candidates are matched to active needs.

Selection and development

Selection includes an English language screening with an AI interviewer about your background, followed by a specialist interview covering indexing, query plans, data modeling, window functions, CTEs, and transaction management. Domain assessments form part of the process.

The work has defined expectations and independent ownership of deliverables. Consistently strong contributors may take responsibility for new programs and mentor incoming subject experts, while helping improve models used by AI labs and major enterprises.