Senior Software Engineer, High Performance Computing
Zilliz
·
Posted:
April 5, 2023
·
Onsite
About the position
Zilliz is a fast-growing startup that develops the industry's leading vector database company for enterprise-grade AI. They are looking for a C++ proficient engineer to optimize core algorithm libraries, create novel algorithms, and participate in the development of high-performance vector indexing frameworks. The ideal candidate should have experience designing and developing large-scale infrastructure software and familiarity with parallel programming and micro-optimization techniques. Zilliz is committed to building an inclusive and diverse workforce.
Responsibilities
- Optimize core algorithm libraries and create novel algorithms for vector preprocessing, index building, and querying.
- Analyze algorithm performance, and formulate technical plans and benchmarks for improvements.
- Participate in the development of high-performance vector indexing frameworks.
- Provide support for the optimization of popular vector indexing algorithms.
- Implement high-performance algorithms from research papers.
Requirements
- Optimize core algorithm libraries and create novel algorithms for vector preprocessing, index building, and querying.
- Analyze algorithm performance, and formulate technical plans and benchmarks for improvements.
- Participate in the development of high-performance vector indexing frameworks.
- Provide support for the optimization of popular vector indexing algorithms.
- Implement high-performance algorithms from research papers.
- Proficiency in C++ is a must.
- Embracing the engineering mentality. Experience designing and developing large-scale infrastructure software such as distributed databases or parallel data processing frameworks is required.
- Familiarity with parallel programming (e.g., OpenMP) and micro-optimization techniques.
- Strong analytical skills and trouble-shooting ability. Strong sense of responsibility and ownership.
- Experience in vector similarity search algorithms or GPU programming (e.g., CUDA) is preferred.