Senior Software Engineer, Vector Index Research

Job description

Zilliz is a fast-growing startup developing the industry’s leading [vector database](https://zilliz.com/) for enterprise-grade AI. Founded by the engineers behind Milvus, the world’s most popular [open-source vector database](https://milvus.io/), the company builds next-generation database technologies to help organizations quickly create AI applications. On a mission to democratize AI, Zilliz is committed to simplifying data management for AI applications and making vector databases accessible to every organization. The Vector Index team focuses on building the core vector retrieval capabilities behind Milvus, Zilliz Cloud, and Vector Lakebase. We work on making similarity search over massive embedding datasets faster, more accurate, and more cost-efficient, while continuously advancing ANN algorithms, index structures, quantization, compression, recall optimization, CPU/GPU acceleration, and high-performance retrieval frameworks. This role sits at the intersection of research and engineering. You will read papers, evaluate new algorithms, build prototypes, and turn promising ideas into production-grade vector indexing and retrieval systems. We are looking for engineers who enjoy research, but also have strong engineering fundamentals, performance optimization skills, and engineering taste. ### What you'll do: - Research, evaluate, and implement new vector indexing and retrieval algorithms for Milvus, Zilliz Cloud, and Vector Lakebase - Read papers and track emerging work in vector search, ANN algorithms, index structures, quantization, compression, reranking, GPU acceleration, and AI retrieval systems - Build high-performance vector indexing components, including index building, query paths, vector preprocessing, quantization, compression, memory layout, and CPU/GPU acceleration - Optimize vector retrieval performance across latency, throughput, recall, memory usage, index build time, and cost efficiency - Design benchmarks and evaluation frameworks to compare algorithms and implementations under real data scale, real query patterns, and real AI workloads - Debug and solve complex performance issues across algorithm implementation, CPU/GPU execution, SIMD/vectorization, memory access, concurrency, and I/O - Turn research prototypes into maintainable, testable, and evolvable production-grade indexing capabilities - Use AI tools across the research and engineering workflow, including paper analysis, prototype generation, code implementation, testing, benchmarking, documentation, and performance analysis ### What we're looking for: - 3+ years of experience in vector search, ANN algorithms, search systems, high-performance computing, or performance-critical systems - Bachelor's degree in Computer Science, Software Engineering, or a related field, or equivalent practical experience - Strong C++ or Rust programming ability and solid engineering fundamentals - Experience with vector similarity search, ANN algorithms, index structures, quantization, compression, reranking, or high-performance retrieval systems is a strong plus - Strong interest in research-driven engineering: reading papers, evaluating tradeoffs, building prototypes, and turning ideas into production systems - Experience with performance optimization and systematic debugging is a strong plus, especially around CPU/GPU execution, SIMD, memory layout, concurrency, I/O, or large-scale data processing - Interest in using AI tools to improve research, coding, testing, benchmarking, documentation, and performance analysis ### How we operate: - **Research-driven, production-focused:** We track frontier algorithms, but care most about whether they work under real data scale, real query patterns, and real production constraints - **Extreme performance:** We care about every memory access, every query path, and every tradeoff between recall and latency - **AI-first engineering:** We actively use AI to accelerate paper reading, prototyping, coding, testing, documentation, and performance analysis, but human judgment and engineering taste still matter most - **Fast and pragmatic:** We work on hard vector indexing and retrieval problems, but we ship them into Milvus, Zilliz Cloud, and Vector Lakebase - **Open source by default:** Milvus is a core part of our engineering culture, and strong indexing capabilities should stand up to public design, code, and community usage ### Benefits: - Competitive compensation (cash + equity) - Regular bonus and equity refresh opportunities - Medical, dental, and vision insurance - Paid time off, including vacation, sick leave, and global reset/wellbeing days - Generous 401(k) and regional retirement plans