Role Overview
Database Search team builds core database-engine capabilities to optimize query performance, including indexing, data pruning, and query optimization. The role focuses on designing and implementing features within the database engine.
Responsibilities
- Implement, test, and improve database-engine features across indexing, data pruning and skipping, query optimization, and execution.
- Investigate performance and scalability bottlenecks using debugging, instrumentation, and measurement, then drive fixes through validation.
- Write maintainable, reliable code and tests in C++ and/or Java.
- Build and improve components in a highly parallel database engine, reasoning about how changes affect query execution end to end.
- Work with teammates and partner engineering groups to refine designs, deliver changes, and share results.
- Evaluate changes against customer workloads and system behavior, and measure their performance impact.
Requirements
- Experience building software through professional work, internships, research, or substantial personal or open-source projects.
- Proficiency in C++ and/or Java, with an interest in writing efficient, reliable systems software.
- Foundational knowledge of data structures, algorithms, and software design.
- Familiarity with SQL and core database concepts, or a strong interest in developing that knowledge.
- Foundational understanding of concurrency and multithreading.
- A proactive, self-motivated approach to hard problems: you take initiative, pursue issues to root cause, and work through ambiguity while communicating progress and incorporating feedback.
- Professional proficiency in English.
- A degree in Computer Science or a related field, or equivalent practical experience.
Helpful Experience
- Exposure to database internals, query engines, storage engines, indexing, or query execution.
- Experience analyzing performance in concurrent, multithreaded, or otherwise performance-sensitive systems software.
- Familiarity with SQL workloads or columnar data formats such as ORC or Parquet.