Contribute to the design, development, and testing of various inference optimization algorithms in the LLM-compressor, Speculators, and vLLM projects.
Design, implement, and optimize model compression pipelines using techniques such as quantization and pruning.
Develop and maintain speculative decoding frameworks to improve inference speed while maintaining model accuracy.
Collaborate closely with research scientists to translate experimental ideas into robust, production-ready systems.
Profile and optimize end-to-end LLM performance, including memory usage, latency, and throughput.
Benchmark, evaluate, and implement strategies for optimal performance on target hardware.
Build tools to streamline model training, evaluation, and deployment.
Participate in technical design discussions and propose innovative solutions to complex problems.
Contribute to open-source projects, code reviews, and documentation; collaborate with internal and external contributors.
Mentor and guide team members, fostering a culture of continuous learning and innovation.
Stay current with LLM architectures, inference optimizations, quantization research, and CPU/GPU hardware advancements.

Strong understanding of machine learning and deep learning fundamentals with experience in one or more of LLM Inference Optimizations and NLP
Experience with tensor math libraries such as PyTorch and NumPy
Strong programming skills with proven experience implementing Python based machine learning solutions
Ability to develop and implement research ideas and algorithms
Experience with mathematical software, especially linear algebra
Understanding of Linear Algebra, Gradients, Probability, and Graph Theory
Strong communications skills with both technical and non-technical team members
BS, or MS in computer science or computer engineering or a related field.
A PhD in a ML related domain is considered a strong plus.

Machine Learning Engineer

Key skills