Micron
New College Grad - Memory System Architect, HBM Generative AI
$103K–$210K
Folsom, CA +1 · Full-time · Posted Oct 13, 2026
- Location
- Folsom, CA
Richardson, TX - Experience
- 0–2 years of relevant experience through research, internships, academic projects, publications, or industry experience.
- Education
- Master's degree, PhD, or equivalent experience in Computer Architecture, Electrical Engineering, Computer Engineering, Computer Science, or a related technical field.
About the job
Research and model future memory architectures for generative AI and data-intensive workloads, informing product strategy and technology roadmaps.
What you'll do
- Research memory architecture concepts for AI accelerators and emerging compute platforms.
- Develop system-level models to assess performance, power, scalability, and efficiency across memory and compute architectures.
- Evaluate memory hierarchy optimizations and architectural trade-offs, and collaborate on recommendations and technology roadmaps.
What they ask for
- Strong understanding of GPU, TPU, AI accelerator, or large-scale computing architectures.
- Knowledge of memory hierarchy concepts, including caches, HBM, DRAM, storage systems, and memory performance optimization.
- Experience with performance modeling, architectural simulation, system analysis, or related research methodologies.
- Programming experience in Python, C++, and/or AI computing frameworks such as CUDA, TensorFlow, or similar technologies.
Nice to have
- Research experience in AI accelerators, computer architecture, memory systems, high-performance computing, or large-scale distributed systems.
- Experience with compute-near-memory, processing-in-memory, or advanced memory architecture concepts.
Pay
$103,000.00 - $210,000.00 USD a year (base salary)
Tools
GPU, TPU, HBM, DRAM, Python, C++, CUDA, TensorFlow
AI focus
Research and evaluate memory architectures and system designs supporting generative AI, AI accelerators, and data-intensive workloads; familiarity with AI-enabled tools and LLMs is also stated.
Model-written summary · the original posting is authoritative