Member of Technical Staff - Tokenomics

Summary

SemiAnalysis is an independent research and analysis firm specializing in the semiconductor and AI industries. The Member of Technical Staff will develop AI inference benchmarks and system models, author technical research reports and newsletter articles, build industry partnerships, and track emerging technologies through conferences.

Responsibilities

• Conduct training & inference performance benchmarks across various AI hardware (e.g. NVIDIA H100, AMD Mi300X, Google TPUs, AWS Trainium2) using frameworks such as PyTorch, JAX, vLLM, SGLang, etc
• Author detailed technical research reports analyzing benchmark results, hardware performance, scalability, & efficiency
• Develop comprehensive system modelling using Python & NCCL for existing & future AI compute clusters, scaling from single-GPU setups to O(100k) GPU clusters
• Establish and maintain strategic partnerships & collaborations with over 50 leading neocloud providers & AI chip manufacturers, including AMD, NVIDIA, and other industry stakeholders
• Stay current on emerging trends & technologies by attending major industry & academic conferences such as NeurIPS, MLSys, NVIDIA GTC, AMD’s Advancing AI, etc

Skills

• Proactive, self-motivated, and capable of working independently in a global team
• Demonstrated experience in ML frameworks such as PyTorch or JAX through professional experience, personal projects, or personal Substack blogs
• Solid understanding of at least 1 of the following: transformer architecture, LLM parallelism strategies, and/or CUDA parallel programming
• Strong research skills and the ability to synthesize information from various sources to draw insights
• English language proficiency required
• Chinese language proficiency preferred

Qualifications

Must Haves

• Proactive, self-motivated, and capable of working independently in a global team
• Demonstrated experience in ML frameworks such as PyTorch or JAX through professional experience, personal projects, or personal Substack blogs
• Solid understanding of at least 1 of the following: transformer architecture, LLM parallelism strategies, and/or CUDA parallel programming
• Strong research skills and the ability to synthesize information from various sources to draw insights
• English language proficiency required

Nice to Haves

• Chinese language proficiency preferred

Benefits

• Paid coding challenge as part of the interview process
• Direct authorship recognition for newsletter articles
• Generous PTO
• Office stipends
• Competitive healthcare (medical, dental, vision)
• Support for conferences and ongoing learning