Senior System Software Engineer, ML and Vector Search
NVIDIA
Opportunity published from Workday · applications stay with the original hiring source.
About the opportunity
NVIDIA is expanding its capabilities into data science, data processing, and database acceleration.
We are seeking an outstanding software engineer to develop our unstructured data processing libraries, NVIDIA cuVS and cuML, which compile highly optimized C++ and CUDA algorithms and expose them to Java and Python.
In this role, you will research, develop, benchmark, and explore novel tuned custom algorithms to accelerate vector preprocessing, indexing, and search, while building high-performance machine learning algorithms for clustering and visualization.
This is a unique opportunity to apply your CUDA and C++ skills to vector search at the intersection of databases and machine learning.
Key responsibilities
- Analyze, design, and implement optimized GPU algorithms for large-scale vector search, databases, and machine learning.
- Drive performance analysis, benchmarking, and optimization of associated libraries.
- Collaborate with a multi-functional team to understand requirements and implement or improve solutions.
- Develop and improve machine learning algorithms.
- Implement solutions in C++, CUDA, and Python.
- Contribute to open source projects like cuML, cuVS, and RAFT.
Qualifications
- BS, MS, or PhD in Computer Science, Data Science or AI, Applied Math, or related field (or equivalent experience).
- Minimum of 8+ years of experience programming in C++ or the ability to learn it by moving from a similar language such as C, Rust, or Java.
- Experience with machine learning concepts, including approximation versus exact methods, iterative versus closed-form approaches, writing ML algorithms, and using ML to solve problems.
- Strong analytical problem-solving skills with a solid foundation in algorithms and mathematics.
- Excellent software development skills covering programming, debugging, performance analysis, and test design.
- Ability to work independently and manage your own development efforts.
- Good communication and documentation habits with a deep care for robust, readable, high-performance code.
- Familiarity with at least one parallel programming or concurrency framework, such as CUDA, OpenMP, OpenACC, Java concurrency, or pthreads.
Nice to have
- Experience developing distributed algorithms and running on distributed systems including HPC and cloud environments.
- Strong debugging skills, including experience with complex multi-language and multi-hardware systems.
- Familiarity with vector databases and nearest-neighbor algorithms such as Milvus, Pinecone, LanceDB, and graph or IVF indexes.
- Machine learning background including clustering, dimensionality reduction, probability, formal methods, and applied research.
- GPU programming background, though GPU development training is provided.
Benefits
- Competitive base salary range from 184,000 USD to 287,500 USD, determined by location, experience, and peer pay.
- Equity eligibility.
- Comprehensive benefits package.
Job details
- Location: Remote / on-site / hybrid (refer to specific regional postings).
- Minimum experience: 8+ years.
- Application deadline: Accepted at least until October 9, 2026.
- Work authorization: Standard US work authorization requirements apply.
About the company
NVIDIA is widely considered one of the technology world's most desirable employers, featuring some of the most forward-thinking and hardworking people in the industry. We are pioneering advancements in accelerated computing, artificial intelligence, data science, and graphics.
Join us to make a huge impact and work with cutting-edge technologies that are transforming industries worldwide.
Additional information
- This posting is for an existing vacancy.
- NVIDIA uses AI tools in its recruiting processes.
Application: UpVagas lists this opportunity from the hiring company or its recruitment platform. The application is completed at the original source.
Full listing
UpVagas condensed repetitive or legal boilerplate to make this page easier to read. The original hiring source contains the complete job description.
