Principal Model Optimization Engineer

Overview Roblox is hiring a Principal Model Optimization Engineer for ML Platform in San Mateo. The role focuses on improving GPU-based machine learning model performance for training and inference, with experience in CUDA, Triton, TensorRT, and LLM optimization techniques. Responsibilities - Optimize machine learning models for GPU architectures - Profile low-level performance bottlenecks in ML pipelines - Develop best practices and tooling for model optimization and deployment - Collaborate with data scientists and software engineers to deploy optimized models - Build tooling, interfaces, and visualizations for internal ML users Requirements - 6+ years of professional experience - Strong GPU debugging experience, including GPU profiles and Xid errors - Experience with CUDA, Triton, and TensorRT - Experience with LLM optimization techniques such as speculative decoding, continuous batching, and quantization - Bachelor's degree in a relevant technical field Benefits - Annual salary range: $295,250–$345,040 USD for San Mateo-based roles - Eligible for equity compensation and benefits - Hybrid office schedule: onsite Tuesday through Thursday, optional Monday and Friday Source https://careers.roblox.com/jobs/7872401?gh_jid=7872401