Job Description:
ScyllaDB is a leading provider of high-performance, highly scalable NoSQL database solutions. Established with a vision to overcome the limitations of traditional databases, we have engineered a robust, open-source database that delivers exceptional speed, low latency, and unmatched reliability. Our mission is to empower businesses to achieve new heights of efficiency, scalability, and agility in managing their data.
The co-founders of ScyllaDB, Dor Laor and Avi Kivity, represent the company’s DNA that is well established in the open source community - originally introducing KVM to the Linux kernel.
The culture in ScyllaDB reflects that openness.
Even as a rapidly growing business, we strive to preserve the qualities of the startup spirit, by keeping management overhead low and encouraging developers to innovate and contribute both as individuals and as part of the whole team, while adhering to the highest standards of quality by producing high quality, well reviewed and tested code, and implementing sophisticated automated system testing, management and support tools.
Team and Challenges
ScyllaDB Core is a group of circa 30 engineers focused on making Scylla Open Source and Scylla Enterprise products great. We work on all areas of the product, including log-structured merge tree storage, the query compiler and execution engine and clustering, i.e. re-balancing the data in and automatically restoring from a degraded state.
ScyllaDB core server is open source. It is built over the high-performance Seastar framework which allows us to get the 10x performance using a shard-per-core architecture with asynchronous, cooperative preemption scheduling and share-nothing memory management. ( tutorial ).
Position
The primary tool we use for writing the test is Python, so it’s best if you know it well and enjoy writing in it. You also need to know and love databases. The kind of work we do is building long-running tests that stress clustered software and reducing them to a specific problem when they fail. Best if both steps are automated. Just like Bubba loved shrimp, we love our tests on a variety of hardware and virtualization platforms, in and out continuous integration, running for a long time and a fraction of a second, and the list goes on. We aim at maintaining a sound test pyramid when test failures are easy to trace to a specific problem, and flaky failures are rare and expect you to be savvy about using computing resources when extending the coverage.
