Memory Architecture Research Intern for AI Systems

1040 Micron Semiconductor Prds; California (MO); Hybrid; USD 69,000 - 77,000; Full time

Job summary

Micron Technology is seeking a Memory & System Architecture Research Intern to explore memory-centric architectures for datacenter and AI workloads, translating workload trends into memory capacity, bandwidth, and latency requirements. You will develop and validate performance models, contribute to publications, and work with senior memory and systems architects on challenging AI infrastructure problems.

Applicants should be pursuing an MS or PhD in Computer Architecture, Computer Engineering,

Qualifications

• Pursuing an MS or PhD in a related field.
• Strong background in computer architecture and system-level performance analysis.
• Proficiency in Python for analytical modeling.
• Experience with memory architectures and AI workloads (preferred).

Responsibilities

• Characterize AI workloads to derive memory bandwidth, capacity, and latency needs.
• Contribute to analytical and simulation models for memory-centric architectures.
• Assess system tradeoffs across memory technologies and integration approaches.
• Benchmark models against reference data to ensure accuracy and reproducibility.
• Produce quantitative analyses to inform architectural decisions.

Skills

Python

Computer architecture

Performance analysis

Analytical thinking

Education

MS or PhD in Computer Architecture / Electrical Engineering

Tools

Simulation frameworks

AI accelerators

Job description

Micron Technology is seeking a Memory & System Architecture Research Intern to explore memory-centric architectures for datacenter and AI workloads, translating workload trends into memory capacity, bandwidth, and latency requirements. You will develop and validate performance models, contribute to publications, and work with senior memory and systems architects on challenging AI infrastructure problems.

Applicants should be pursuing an MS or PhD in Computer Architecture, Computer Engineering,