Machine Learning Research Engineer
Core
Researching and optimizing AI model architectures and inference algorithms specifically for Etched's custom ASIC hardware (Sohu) to achieve performance unattainable on GPUs.
Role type
Machine Learning Research Engineer (HW Co-design)
Builds
Custom AI chips (ASICs) and optimized inference software stacks for generative AI models.
Domain
Semiconductor hardware engineering and Large Language Model (LLM) inference optimization.
Deliverable
production ML models | infrastructure
Required skills
Transformer model architectures, Python, PyTorch, JAX, distributed inference/training environments, model-specific inference-time acceleration techniques.
Preferred skills
ML Systems Research, HW Co-design, Rust, GPU kernels, CUDA compilation stack, published inference-time compute research.
Technologies
Python, Rust, PyTorch, JAX, vLLM, SGLang, Sohu ASIC.
Responsibilities
Propose and conduct novel research to achieve results on Sohu that are unviable on GPUs; Translate core mathematical operations from Transformer-based models into maximally performant instruction sequences for Sohu; Develop deep architectural knowledge informing best-in-the-world software performance on Sohu HW; Co-design and finetune emerging model architectures for highest efficiency on Sohu; Implement frontier models using Python and Rust to guide and contribute to the Sohu software stack.
Seniority
Mid-to-Senior, hands-on IC
