Hardware-Aware Model Optimisation - Member of Technical Staff
Core
Develop hardware-aware transformations and automated methods to adapt LLM architectures and inference algorithms for heterogeneous accelerators, optimizing trade-offs between capability, latency, and cost.
Role type
Senior IC machine-learning engineer (hardware-aware optimization)
Builds
Hardware-specialized model variants and automated optimization pipelines
Domain
AI infrastructure / Heterogeneous computing / LLMs
Deliverable
production ML models
Required skills
LLM architecture design, efficient inference, hardware-aware machine learning, numerical/structural/memory optimization, experimental design
Preferred skills
Mixture-of-experts, diffusion models, JEPA, speculative decoding
Technologies
Cerebras, d-Matrix, Intel, NVIDIA, AMD, Normal Computing, Tenstorrent, GreatSky, Mixx
Responsibilities
Develop hardware-aware transformations for model adaptation; Explore numerical and memory-related optimizations; Investigate new inference algorithms; Build automated discovery and optimization methods; Establish evaluation metrics for hardware adaptations
Seniority
Senior, hands-on IC