Hire Collin H. - Forward Deployed Engineer

* Designing and maintaining distributed and componentized tools for data scientists to build ML models at scale using Python, Dask, Kubeflow, and XGBoost. * Achieved a 40% reduction in time on a 50% increase in data in a model fit pipeline by optimizing Dask configurations and usage * Standardized repo and engineering practices to provide consistency between repos * Incorporated (and documented incorporation strategies of) better environment manager into components while abiding by proprietary and strict CI/CD processes, accelerating local development by 100x in some cases

Designing and maintaining distributed and componentized tools for data scientists to build ML models at scale using Python, Dask, Kubeflow, and XGBoost. - Reduced KFP (Kubeflow Pipelines) pipeline.yaml size by 13.7% for Capital One's largest Kubeflow pipeline, allowing engineers flexibility to develop new features without hitting the pipeline file limit size enforced by KFP. - Created a reusable recursive feature eliminator Kubeflow component, unifying customers and streamlining support under a single component - Parallelized a data migration workflow to increase transfer speed from months to days, enabling the decommissioning of an old expensive Kubernetes cluster, ultimately saving $50,000+