Lyon College ’25 | Computer Science & Data Science | ML, analytics & AI engineering
Data scientist · ML practitioner · AI engineer
I turn messy data into decisions people can act on. My core toolkit is classic ML and statistics — regression, tree ensembles, careful cross-validation — with growing depth in AI engineering: LLM-backed tools, prompt and workflow design, retrieval, and evaluation harnesses that keep models honest in production.
I picked up problem-solving the long way: six years working alongside my dad at Alpine Appliance diagnosing washers, dryers, and refrigerators before I ever wrote a Python script. That habit — observe, hypothesize, test, swap a part, repeat — is the same loop I use on a stubborn model now.
I’m looking for roles where I can ship: data scientist, ML engineer, or AI engineer. Remote or relocate — comfortable owning a project end to end.
Lyon College · 2021–2025
Alpine Appliance · 2018–present (intermittent)
Field appliance-repair technician; later refurbishing and reselling used appliances. Built the habits — root-cause diagnosis, working with documentation, telling a customer the truth — that I carry into every codebase.
Featured work, then what’s next on deck.
Tools: R, GBM, XGBoost, kNN, Random Forest
Goal: Automate Comparative Market Analysis — predict sale prices from historical comps and property features.
Result: Optimized blend (15% XGBoost / 85% GBM) — MAE 6,776 · RMSE 11,166
Tools: Python, Azure Document Intelligence, pandas, OCR
Goal: Convert scanned signature sheets into a structured CSV for downstream analysis.
Result: Printed fields (name, address, county) extracted cleanly; handwritten signatures flagged for human review.
Tools: OpenAI / Anthropic APIs, LangChain, pgvector, eval harness
A small set of LLM demos — a RAG pipeline, a structured-output extractor, and an eval suite that fails loudly when prompts drift. Notebook + repo link landing here soon.
Tools: Python, Selenium, BeautifulSoup, Scrapy, pandas
Compare building-supply prices across local stores (2x4s, drywall, paint). Proof of concept handles HTML snapshots and dynamic content; migrating crawl to Scrapy.
Tools: Python, Flask, Google Sheets API, Render
Help local stores without online catalogs surface real-time prices on a searchable website by bridging their inventory system. In talks with a store owner for API access.
Languages and toolkits behind the projects above.
3+ years across research, coursework, and shipping projects. My default language for data work and back-end glue.
Tools: pandas, NumPy, scikit-learn, PyTorch, FastAPI, Selenium, BeautifulSoup
My go-to for statistical modeling and quick visualization. Used heavily on ProphetCMA for ensemble tuning.
Tools: tidyverse, ggplot2, caret, xgboost, gbm, randomForest
Classical ML across regression, tree ensembles, kNN, and blending. Comfortable with cross-validation, leakage checks, and honest metric reporting.
Tools: scikit-learn, XGBoost, GBM, Random Forest, Optuna
LLM-backed tools end to end: API integration, prompt and workflow design, retrieval, and evaluation that catches regressions before users do.
Tools: OpenAI / Anthropic SDKs, LangChain, vector DBs (pgvector / Chroma), Pydantic, eval harnesses
Window functions, CTEs, query planning. Comfortable in Postgres and SQLite; happy living in the query plan when something is slow.
Tools: PostgreSQL, SQLite, pgvector, sqlalchemy
6+ years; my first serious language. Used for games, small apps, and OOP-heavy coursework.
Tools: JavaFX, JUnit, Maven, LWJGL
4+ years building front-ends. This site is hand-rolled HTML + Tailwind; comfortable picking up frameworks when a project demands them.
Tools: Vanilla JS, Tailwind CSS, React (when needed), Vite
2 years through coursework. Useful when I need to read library internals or care about memory layout.
Tools: STL, CMake, gdb
For hiring teams and introductions, you can download my cover letter as a PDF. You can also reach me directly in the contact section below.
Download cover letter (PDF)