ML Research Scientist - Inference ML
Salary not listed • United States and Canada • Full-time
Posted 28d ago
No salary is listed. Many US states require employers to post a pay range, so ask about pay early.
Job Location
United States and Canada
Tech Stack
Remote Work Policy
On-site
Employment Type
Full-time
Categories
LLM Engineer
About the job
Cerebras Systems is seeking an ML Research Scientist to join their Inference ML team. This role involves inventing and evaluating advanced algorithms, methods, and model architectures for efficient inference on Cerebras' flagship AI hardware. You will collaborate with researchers and engineers to design, prototype, validate, and optimize cutting-edge inference techniques, focusing on areas like speculative decoding, model pruning and compression, sparse attention, and sparsity-driven methods to achieve low-latency, high-throughput inference at scale. The position offers the opportunity to push the boundaries of AI inference and contribute to the development of the world's fastest AI accelerator.
Responsibilities
- Research and prototype novel inference algorithms and model architectures leveraging Cerebras hardware, focusing on speculative decoding, pruning/compression, sparse attention, and sparsity.
- Design and evaluate new transformer architectures for NLP and computer vision to optimize quality, latency, throughput, and compute trade-offs.
- Identify and pursue high-impact research directions at the intersection of model architecture, inference systems, and hardware.
- Train or adapt models and analyze inference performance using Cerebras tools to maximize throughput and minimize latency.
- Collaborate with engineering and performance teams to translate research into usable systems.
- Publish and present research findings and contribute to open-source releases.
Requirements
- PhD in machine learning, computer science, or related field with 3+ years of relevant research/industry experience, OR Master's degree with exceptional research record and 4+ years of relevant experience.
- Deep understanding of modern ML architectures, generative models (language/vision), ML systems, and inference behavior.
- Ability to reason about model and algorithm performance for inference workloads.
- Ability to identify relevant research questions, develop novel methods, and evaluate their impact on model quality and inference efficiency.
- Evidence of research impact through publications, preprints, or open-source contributions.
- Strong Python programming skills and proficiency with ML frameworks like PyTorch, Transformers, vLLM, or SGLang.
- Original research in speculative decoding, neural network pruning/compression, sparse attention, quantization, mixture-of-experts, sparsity, post-training techniques, or inference-aware design (preferred).
- Track record of owning complex ML or inference projects end-to-end (preferred).
- Experience designing novel model architectures (preferred).
- Strong foundation in performance optimization on specialized hardware (e.g., GPUs, TPUs, HPC interconnects) (preferred).
- Familiarity with large-scale model training and deployment, including performance and cost trade-offs (preferred).
- Experience with Triton, CUDA, or C++ (a plus).
Benefits
- Build a breakthrough AI platform beyond the constraints of the GPU.
- Publish and open source cutting-edge AI research.
- Work on one of the fastest AI supercomputers in the world.
- Job stability with startup vitality.
- Simple, non-corporate work culture that respects individual beliefs.