About The Role
As a Research Engineer on the Inference ML team at Cerebras Systems, you will adapt today's most advanced language and vision models to run efficiently on our flagship Cerebras architecture. You'll work alongside ML researchers and engineers to design, prototype, validate, and optimize models, gaining end-to-end exposure to cutting-edge inference research on the world's fastest AI accelerator.
You will focus on pushing the frontier of speculative decoding, large-model pruning and compression, sparse attention, and sparsity-driven techniques to deliver low-latency, high-throughput inference at scale.
Responsibilities
- Implement and adapt transformer-based models (NLP and/or vision) to run on Cerebras hardware
- Assist in optimizing models for inference performance (latency, throughput)
- Run experiments, analyze results, and support model improvements
- Help bring up and validate models on the Cerebras system
- Debug and troubleshoot model or system issues with guidance from senior team members
- Support profiling and performance analysis using internal tools
- Collaborate with cross-functional teams (ML, software, hardware) on model integration
Minimum Qualifications
- Bachelor’s or Master’s degree in Computer Science, Engineering, or a related field
- 1–3 years of experience in software engineering or machine learning in a similar capacity (internships count)
- Experience with Python and at least one ML framework (e.g., PyTorch, Transformers, vLLM or SGLang)
- Understanding of deep learning concepts (e.g., neural networks, transformers)
- Experience with Generative AI and Machine Learning systems
- Strong programming skills in Python and/or C++
Preferred Qualifications
- Experience with speculative decoding, neural network pruning and compression, sparse attention, quantization, sparsity, post-training techniques, and inference-focused evaluations.
- Exposure to large language models or computer vision models
- Experience running experiments or tuning models
- Familiarity with tools like PyTorch, Hugging Face Transformers, or similar
- Basic understanding of performance concepts (e.g., latency, throughput)
- Experience working in Linux environments