Database Systems Expert - PhD

Mercor; Berlin; On-site; EUR 96,000 - 165,000; Full time

Job summary

Obsidian is seeking a senior database engineer to design and validate complex, self-contained tasks that test AI agents on database internals. Tasks cover storage engines, transactions, replication, and low-latency considerations.

Applicants should have 5+ years of hands-on experience or a database-systems PhD, and must clearly articulate correctness and performance tradeoffs. A paid task (~25 hours) follows selection.

Qualifications

• Minimum five years of hands-on database-engine internals experience or a database-systems PhD.
• Deep experience with storage engines, query execution, optimization, transactions, replication, or related low-latency systems.
• Strong implementation skills and ability to explain correctness and performance tradeoffs clearly.
• DBA, SQL analytics, ETL, or application-only database experience does not qualify on its own.

Responsibilities

• Design one challenging, self-contained database engineering task.
• Build checks for correctness, failure handling, concurrency, retries, and regressions where relevant.
• Evaluate performance measures such as latency, throughput, and memory use.
• Write clear grading criteria for engineering judgment.
• Work with technical reviewers to refine and validate the task.

Skills

Database internals

Storage engines

Query execution

Optimization

Transactions

Replication

Education

PhD in database systems

Tools

Python 3.8+

C

C++

Rust

Go

Java

Job description

Applications close Tuesday, October 13, 2026 at 12:00 p.m. Pacific. We are selecting the first three applicants who qualify for the initial pilot.

About the work

Build real-world tasks that test how well AI agents handle database engineering. Depending on your specialty, your task may cover database-engine implementation, production diagnosis, or application migration.

Each task should include a clear problem statement, a reproducible repository or environment, a working reference solution, tests, and grading criteria that distinguish a robust engineering solution from one that only appears correct.

What you'll do

• Design one challenging, self-contained database engineering task.
• Build checks for correctness, failure handling, concurrency, retries, and regressions where relevant.
• Evaluate performance measures such as latency, throughput, and memory use.
• Write clear grading criteria for engineering judgment.
• Work with technical reviewers to refine and validate the task.

Who we're looking for

• At least five years of hands-on database-engine internals experience, or a database-systems PhD.
• Deep experience with storage engines, query execution, optimization, transactions, replication, or related low-latency systems.
• Strong implementation skills and clear explanations of correctness and performance tradeoffs.

DBA, SQL analytics, ETL, or application-only database experience does not qualify on its own. You do not need to specialize in every task area.

Application process

Complete a self-paced technical assessment with written responses and a coding exercise. You control when to advance between sections. (~35 min)

Use a laptop or desktop with a camera, microphone, and entire-screen sharing. Have your own editor and terminal ready, along with Python 3.8+ and a compiler for C, C++, Rust, Go, or Java. The assessment includes a starter kit and public tests. Complete the assessment independently, without outside AI assistance.

Paid task after selection

Selected experts will complete one task, with approximately 25 hours of work, followed by technical review and feedback.

Task deadline: Friday, October 30, 2026.