Senior SDE: Neuron Inference for LLM Acceleration

Amazon Inc.; Seattle (WA); On-site; USD 144,000 - 194,000; Full time

Benefits offered by this job

Health insurance

401(k) matching

RSUs

Job summary

Amazon AWS Neuron is seeking a senior software engineer for the Machine Learning Inference Applications team in Seattle. You will develop and optimize core building blocks of LLM inference on Neuron chips, including attention, MLP, quantization, speculative decoding, and Mixture of Experts.

Responsibilities include applying the latest research in LLM optimization to extract top performance from open source and internal models, and working across teams to deliver scalable, high-performance

Qualifications

• 3+ years of non-internship professional software development experience.
• 2+ years of non-internship design or architecture of systems.
• 1+ years of SDE or related experience.
• 1+ years of designing large-scale multi-tier distributed software.
• 1+ years of OO design experience.
• Bachelor's degree or foreign equivalent in CS, Engineering, Math, or related field.
• Experience programming with at least one language (C#, C++, Java, or Perl).

Responsibilities

• Develop and optimize core building blocks of LLM Inference (Attention, MLP, Quantization).
• Adapt latest research in LLM optimization for Neuron chips to maximize performance.
• Collaborate across teams to deliver scalable software for machine learning inference.

Skills

Software development

System design

OO design

C/C++/Java

Distributed systems

Education

Bachelor's degree in CS or related field

Job description

Amazon AWS Neuron is seeking a senior software engineer for the Machine Learning Inference Applications team in Seattle. You will develop and optimize core building blocks of LLM inference on Neuron chips, including attention, MLP, quantization, speculative decoding, and Mixture of Experts.

Responsibilities include applying the latest research in LLM optimization to extract top performance from open source and internal models, and working across teams to deliver scalable, high-performance