AI Model Evaluation Data Engineer

Obsidian; San Francisco (CA); On-site; USD 207,000 - 551,000; Part time

Job summary

Obsidian is partnering with an AI research initiative to support a Frontier Code Agents project in San Francisco. You will evaluate frontier AI coding models by performing realistic data engineering tasks, reviewing ETL pipelines, data warehouses, and distributed systems.

Join a sprint-based program where each accepted task is compensated, with rapid ramp-up and tight feedback cycles. Candidates should have 2+ years in data engineering and comfort using AI coding agents to assess infrastructure.

Qualifications

• 2+ years of professional data engineering experience.
• Experience building ETL pipelines, data warehouses, analytics platforms, or distributed data systems.
• Regular use of AI coding agents such as Cursor, Claude Code, Codex, Windsurf, Gemini CLI, or similar tools.
• Ability to evaluate model-generated data infrastructure and pipeline implementations.
• Experience operating large-scale data platforms is preferred.

Responsibilities

• Use frontier AI coding agents to complete and evaluate complex data engineering tasks.
• Review model-generated implementations involving ETL pipelines, data warehouses, analytics platforms, and distributed data systems.
• Identify bugs, edge cases, scalability issues, and failure modes.
• Compare outputs from multiple frontier models and assess their strengths and weaknesses.
• Apply professional engineering judgment to realistic data engineering scenarios.

Skills

Data engineering

ETL pipelines

Data warehouses

Analytics platforms

Distributed data systems

AI coding agents usage

Tools

Cursor

Claude Code

Codex

Windsurf

Gemini CLI

Job description

Obsidian is partnering with an AI research initiative to support a Frontier Code Agents project in San Francisco. You will evaluate frontier AI coding models by performing realistic data engineering tasks, reviewing ETL pipelines, data warehouses, and distributed systems.

Join a sprint-based program where each accepted task is compensated, with rapid ramp-up and tight feedback cycles. Candidates should have 2+ years in data engineering and comfort using AI coding agents to assess infrastructure.