About the Role
Design and implement high-performance mathematical libraries that exploit sparsity patterns in AI workloads to improve computational efficiency and reduce memory usage.
Responsibilities
- Develop optimized linear algebra routines tailored for sparse data structures
- Collaborate with AI researchers to identify sparsity opportunities in neural networks
- Improve performance of machine learning frameworks through low-level library optimizations
- Profile and benchmark numerical kernels on GPU architectures
- Contribute to open-source projects related to sparse computation
- Analyze algorithmic complexity of sparse matrix operations
- Work closely with hardware teams to align software with architectural advancements
- Document technical designs and performance results
- Support integration of math libraries into deep learning platforms
- Optimize memory access patterns for sparse tensor computations
- Ensure numerical accuracy and stability in sparse implementations
- Drive adoption of sparsity techniques across AI applications
- Evaluate emerging AI models for sparse computation applicability
- Develop tools for sparsity pattern analysis and transformation
- Contribute to API design for sparse math libraries
- Troubleshoot performance bottlene游戏副本s in real-world AI deployments
- Participate in code reviews and maintain high software quality standards
- Stay current with advancements in sparse numerical methods
- Collaborate on cross-team initiatives involving AI acceleration
- Provide technical guidance to junior engineers
Nice to Have
- PhD in a technical discipline with focus on numerical computing
- Contributions to open-source machine learning projects
- Experience with tensor compilers
- Background in sparse solver development
- Familiarity with AI model compression techniques
- Prior work on math libraries such as cuBLAS or cuSPARSE
- Experience with automated code generation for math kernels
- Knowledge of mixed-precision computing in AI
- Research publications in numerical methods or AI acceleration
- Experience with formal verification of numerical code
Compensation
Competitive salary and benefits package
Work Arrangement
Hybrid work model with flexible remote options
Team
Part of a high-performance computing team focused on AI acceleration
About the Team
This role is part of a specialized group focused on advancing AI performance through mathematical innovation and tight hardware-software co-design.
What We Offer
- Opportunity to work on cutting-edge AI acceleration technologies
- Access to state-of-the-art GPU computing platforms
- Collaborative environment with leading experts in numerical computing and AI
Available for qualified candidates
