Senior Software Engineer, Vector Index Research

Zilliz is building the next generation of data infrastructure for AI. Founded in 2017 by the creators of Milvus, one of the world's most popular open-source vector databases, we help developers and enterprises build, deploy, and scale AI-powered applications, from semantic search and RAG to emerging agentic AI workloads. 

 

Backed by $113 million in funding from leading global investors, Zilliz has established a strong presence in the AI infrastructure market. Milvus has earned over 46,000 GitHub stars, and our technology has reached more than 10,000 enterprise users worldwide. Companies including OpenEvidence, Exa, and Filevine rely on Zilliz Cloud to power their AI applications.

 

Building on our roots in vector database technology, we're expanding Zilliz Cloud into a broader AI data platform with Zilliz Vector Lakebase. As AI continues to reshape how applications are built and how data is used, we're focused on making the underlying infrastructure more powerful, scalable, and accessible to organizations worldwide.

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.

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

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

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

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