Sandisk
Principal AI/HPC Software Engineer
On-site · Milpitas, CA
About the role
Essential Duties and Responsibilities:
AI/HPC Software Engineer
The engineer develops system software, infrastructure software, and AI platform technologies for next-generation AI memory systems. The role sits at the intersection of:
- AI infrastructure
- System software
- Memory systems
- AI/HPC platforms
- Hardware/software co-design
- Energy-efficient AI computing
The goal is to demonstrate and optimize future memory-centric architectures that leverage NAND, storage systems, and emerging AI memory technologies.
Core Responsibilities
AI Infrastructure Software Development
Develop system software and prototype platforms for AI/ML infrastructure:
- AI inference platforms
- AI training environments
- Memory-aware computing platforms
- Distributed AI systems
- Open-source AI infrastructure
The role specifically calls out using computer systems and open-source software technologies to demonstrate next-generation memory architectures.
Memory-Centric Software Architecture
Design software supporting:
- Memory management
- Cache hierarchy optimization
- Data movement protocols
- Storage-memory integration
- AI system runtime environments
The engineer develops software that exposes the advantages of future memory technologies within AI systems.
Hardware/Software Co-Design
Work with architecture teams to:
- Evaluate new memory technologies
- Prototype future AI systems
- Optimize performance and power efficiency
- Validate next-generation AI memory concepts
The role explicitly requires close collaboration with hardware architects.
Frameworks and APIs
Develop:
- Infrastructure software frameworks
- Runtime environments
- APIs
- System software libraries
- Benchmarking and validation tools
for future AI systems.
Required:
Programming
- Python
- C
- C++
Systems
- Operating systems
- Device drivers
- System software design
- System programming
AI Computing
- GPU architecture
- CUDA programming
- AI infrastructure
- Computer architecture
- Memory systems
AI Lab Software Engineer
The strongest candidate would have:
AI Serving
- vLLM; TensorRT-LLM; Triton; SGLang; Dynamo
Distributed AI
- Ray; Kubernetes; Docker; Slurm
GPU Systems
- CUDA; NCCL
- Multi-GPU programming
High-Speed Infrastructure
- RDMA; RoCE
- InfiniBand
Memory-System Research
- HBF
- KV-cache hierarchy
- Storage-memory integration
- AI SSD architectures
Research Capabilities
- Workload tracing
- System simulation
- Digital twins
- AI infrastructure modeling
The AI Lab focused on HBF, KV-cache offload, AI SSDs, and memory-system innovation, this software engineer would essentially function as a LLM Infrastructure + AI Systems Research Engineer, rather than a traditional application software developer.