Micron India Research Center is a part of Technology Pathfinding and Strategy group at Micron India. The goal is to contribute to Micron’s pathfinding in multiple areas like Devices and circuits, Systems and AI-ML with enhanced multi-functional collaboration.

Role and Responsibilities:

• Architect, simulate, and design accelerator architecture. You will interact with multi-functional engineering teams to find opportunities and requirements. Some of the job duties include:

• Implement performance, functional and power model for accelerator architecture exploration and functional verification. Accelerators may include Analog AI using in-memory computing with non-volatile memory technologies.

• Develop tests, test plans, and testing infrastructure to verify new algorithm and functionality and their performance.

• Assess new SW/HW innovations and drive the most relevant into the accelerator Architecture definition.

• Research the industry trend and customer requirement of the machine learning algorithms and accelerators and drive partnerships for access to the most advanced technologies.

Qualification and Skills

• Minimum 12+ years of experience with bachelor’s degree in CS/EE/Math or equivalent. Masters/PhD degree is a big plus.

• 5+ years of Experience with performance simulation and modeling in modern C++/python. Ability to deliver production quality and well-documented code.

• Experience in modeling and benchmarking machine learning models on accelerator simulator models.

• Experience in machine learning frameworks and interfaces, for example: ONNX, TensorFlow and PyTorch. PyTorch is a big plus.

• Knowledge/Experience in CPU/GPU/Domain Specific Accelerators architecture with a focus on data-path, fixed-point arithmetic, scheduler, memory system, and network-on-chip interconnection design

• Experiences on NAND Flash Memory and its architecture and circuit design is a big plus.

• Experiences on DRAM and its architecture and circuit design is a good plus.

• Experience in approximate compute, analog compute is a good plus.

• Experience of hardware and RTL implementation (Verilog/System Verilog) is a good plus.

• Experiences in OpenCL/CUDA/Metal or other such parallel programming APIs is a good plus.

• Passionate in new technology and innovation with strong communication skills. See Less