Sanger Steel
Senior Machine Learning Engineer at CoreWeave
Tampa, Florida, United States
Summary
🤩
Rockstar🎓
Top SchoolSanger Steel is a Senior Machine Learning Engineer based in Tampa with six years of experience building production ML systems and three years focused on leveraging LLMs and NLP for business impact. He’s delivered end-to-end MLOps and model-serving infrastructure at startups and at CoreWeave, and contributed to high-performance open-source tooling by integrating Tensorizer into the vLLM inference engine to speed model loading. Comfortable across research and engineering, he has applied transformer architectures to tasks from classification and NER to summarization and generation while leading small teams to productionize generative AI. A physics-trained problem solver, he also writes about ML (and occasionally physics) on his blog, reflecting a curiosity that bridges rigorous research and pragmatic deployment.
6 years of coding experience
5 years of employment as a software developer
Bachelor's degree Physics, Bachelor's degree Physics at University of St Andrews
Spanish, English
Stackoverflow
51reputation
3kreached
0answers
3questions
Github Skills (10)
pytorch10
llm10
python9
model-management9
inference9
google-cloud-storage6
kubeflow-pipelines6
google-cloud-ml6
google-cloud-functions6
google-cloud-platform6
Programming languages (3)
Github contributions (5)
vllm-project/vllm
Jan 2024 - Sep 2024
A high-throughput and memory-efficient inference and serving engine for LLMs
Role in this project:
ML Engineer
ML EngineerContributions:38 reviews, 6 PRs, 76 comments in 7 months
Contributions summary:Sanger primarily contributed to the integration of Tensorizer, a serialization and deserialization tool, into the vLLM project. Their work involved adding model loading capabilities using Tensorizer, updating documentation, and fixing related test failures. This focused on improving the performance of loading and using models within the vLLM framework. The user also made changes to support automatic detection of vLLM-tensorized models.
amdcudadeepseekgpthpu
coreweave/vllm
Jan 2024 - Feb 2025
A high-throughput and memory-efficient inference and serving engine for LLMs
Contributions:18 reviews, 4 PRs, 281 pushes in 1 year 1 month
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.