Working Student - LLM Inference Performance
Core
Optimize low-level kernels and build profiling tools to improve LLM inference performance on custom silicon hardware.
Role type
Working Student, LLM Inference Performance Engineer
Builds
Low-level inference kernels, benchmarking harnesses, and profiling tools for generative AI workloads
Domain
AI Hardware / Custom Silicon / LLM Inference
Deliverable
production ML models
Required skills
Rust, Python, C/C++, Computer Architecture, Memory Systems, Parallel Computing, Low-level Programming, Kernel Optimization, AI Coding Agents
Preferred skills
GPU/AI Accelerator Programming, Embedded Systems, Performance Profiling, LLM Inference Fundamentals
Technologies
Rust, Python, C, C++, Custom Silicon, AI Coding Agents
Responsibilities
Implement and tune low-level kernels for LLM inference; Build analytical and roofline performance models; Develop benchmarking and profiling tools; Collaborate with compiler and runtime engineers; Evaluate AI coding agents for development workflows
Seniority
Student (Master's level)
