Polymer chemist and materials science consultant specializing in the design, formulation, and scale-up of high-performance chemical products. I’ve worked across Fortune 500 companies and early-stage startups, helping move materials from lab concepts to production in automotive and industrial manufacturing.
My background spans adhesive formulation, surface chemistry, polymer characterization, and materials spec-in for assembly-line production, with hands-on experience across the full product lifecycle—from synthesis and testing to scale-up, customer trials, and launch.
I hold a Ph.D. in Chemistry (University of Illinois Urbana-Champaign) and a B.S. in Chemistry (University of Washington). After 15+ years in experimental R&D, I’m now completing an M.S. in Artificial Intelligence at Seattle University, applying AI/ML to accelerate materials discovery and development.
These repositories are a mix of exploratory projects and anonymized or generalized examples inspired by prior client and industry work.
Local Retrieval Augmented Generation (RAG) for scientific, engineering, and technical research
- Researchers can privately answer search queries, grounded in actual document context, without sharing documents with 3rd parties.
- Supports keyword and semantic search for general and niche queries, previews results for retrieval and prompt tuning.
- Supports text and image fields in corpus documents for comprehensive responses.
- Supports local LLMs for response synthesis and question answering, results and settings can be exported for record keeping, reproducibility, model comparison, and regression testing.
- Workflow for deterministic retrieval, selection, and summarization of high-volume news sources for current awareness.
- API-based scraping, parsing, and LLM-based text summarization, results delivered by email and stored for troubleshooting.
- Supports use as a stand-alone scripted job, scheduled job, or agent skill for convenience.
- Can be tailored to technical or trade publications for team knowledge management.
- Supervised machine learning models for prediction of material indicators for thermal, mechanical, and barrier properties.
- Benchmarks models and hyperparameters in a systematic and reproduceable workflow for data-driven model selection.
- Useful for screening candidates for expensive computational vetting or laboratory synthesis.