Cambridge Research Internship on Compiler-like Tools for LLM Efficiency

1 Day ago • Upto 1 Years • Research & Development

About the job

2 skills required for this role

Add these skills to join the top 1% applicants for this job

pytorch

cuda

Job Description

This internship at Microsoft Research Cambridge focuses on developing compiler-like tools to enhance the efficiency of Large Language Models (LLMs). The intern will research and develop novel tools addressing quantization and parallelization techniques. Responsibilities include designing, implementing, and evaluating these tools, and presenting findings in technical documents and presentations. The ideal candidate will possess substantial experience in static analysis, verification, constraint solving, and working with LLMs, inference, parallelization, and quantization. Proficiency in PyTorch and CUDA is preferred. The internship requires a physical presence in Cambridge, UK during winter/spring 2025.

Must have:

Masters/PhD in CS/ML
Static analysis, verification, constraint solving experience
LLMs, inference, parallelization, quantization experience

Good to have:

Strong technical communication
PyTorch, CUDA, tooling experience

Perks:

Industry leading healthcare
Educational resources
Discounts on products and services
Savings and investments
Maternity and paternity leave
Generous time away
Giving programs
Networking opportunities

Overview

The Future AI Infrastructure team at Microsoft Research Cambridge (UK) is dedicated to developing technologies for AI data centers of the future. We are seeking a masters/PhD student to join us in Cambridge winter/spring 2025 to work on automatic, compiler-like tools for LLM efficiency, covering topics such as quantization and parallelization. You will joining a welcoming and highly collaborative environment and work on creative and challenging problems during your internship.