Skills
Languages, frameworks, and tools — with context on how each was developed and where it shows up in real work.
Languages
Primary language across all CoorsTek work — ML pipelines, data engineering, deep learning training, and app development.
Queried R&D databases at CoorsTek to calculate material formulations and retrieve structured data.
ML & Data Science
Trained semantic segmentation models on XCT and optical images of ceramic parts for geometric analysis. Performed image augmentation to build viable datasets from minimal real data.
Developed and improved ML models for manufacturing quality control at CoorsTek, improving model specificity and F1 from ~80% to ~98%.
Built production CV pipelines for ceramic part analysis — semantic segmentation for internal feature inspection and classification-based defect detection for manufacturing quality control.
Designed and deployed CoorsTek's first internal agentic RAG system from scratch — Databricks vector search, orchestrated AI agents, and custom prompting to reliably answer R&D engineer queries. Established LLMOps evaluation SOPs for the company.
Data Engineering & Infrastructure
Platform for LLM/RAG development and deployment at CoorsTek. Built vector search data pipelines, orchestrated AI agents, and pioneered the company's first Databricks-based LLMOps practices.
Built data ingestion pipelines for structured R&D databases and hundreds of technical documents to power the agentic RAG system.
Software Engineering
Designed a generic OOP implementation of a directed acyclic graph to represent and traverse complex material histories, with methods to calculate formulations and transform units.
Developed a directed acyclic graph (DAG) data structure to model material histories and compute formulation values at CoorsTek.
Built Streamlit web apps for R&D researchers at CoorsTek — enabling material history queries and pixel-level geometric measurements of ceramic parts.
Project & Process
Scrum master for 2 sprints on a cross-functional team of 5; helped instill Agile development practices across the CoorsTek ML team.
Established standard procedures for DataOps, MLOps, and LLMOps across CoorsTek's production-deployed AI products — including agent evaluation and deployment workflows.
Research & Writing
Published writing on IP governance and environmental economics — fitting structured arguments, analysis, and academic citations within tight page limits.
Built and analyzed economic models of waste management policy, synthesizing literature into a novel contribution in Environmental Economics — a model that grew into a peer-reviewed research paper.