Welcome to Braedyn Thompson's portfolio!
See the roles I'm targeting →

Now playing
Berkeley Lab SeismicSoCal BearLM
PLAYER 1 · PORTFOLIO

Braedyn
Thompson

I build machine-learning systems end to end: the data pipeline, the model, the evaluation, and the product people use.

Data engineeringMachine learningLLMs / RAGMLOps
SYSTEM CHECK . . . OK0.9998quake detection ROC-AUC68MBART trip rows modeled48%lower forecast error vs naive82%RAG recall@1 (from 56%)

Roles I'm targeting08

The roles below are where I want to work next. Each one links to the projects and internships that show the skills it needs.

DE

Data Engineer

See the work →PDF ↓

Build pipelines that ingest messy sources reliably and reproducibly.

DA

Data Analyst

See the work →PDF ↓

Turn raw logs and metrics into clear charts and decisions.

ML

Machine Learning Engineer

See the work →PDF ↓

Train models that beat a real baseline, then run them in production.

OPS

MLOps Engineer

See the work →PDF ↓

Make models retrain, promote and monitor themselves, with the gates written down.

SE

Software Engineer

See the work →PDF ↓

Ship full products: backend services, APIs and desktop apps, tested and deployed.

Featured projects06

Experience08

About

I like problems where I can check whether the answer is right.

That check is the first thing I set up. Before I build, I work out what a good answer would look like and how I would know I had one: a baseline worth beating, a held-out set the model never saw, and a number I can stand behind. Most of what follows is there to make that number trustworthy.

My projects run across forecasting, seismology, retrieval and ed-tech, and they follow the same loop. Build a reproducible pipeline, pick a baseline, test on data the model has not seen (chronological splits, ablations, placebo tests), then ship the result as something a person can open and use.

I care about the pipeline as much as the model, because a forecast is only as good as the data under it. So I spend real time on the unglamorous parts: ingestion, cleaning with a logged reason for every dropped row, dbt models with a test on every build, and orchestration that reruns itself on a free VM while I am asleep.

Sometimes the honest answer is that the data cannot support the question. At CBU I spent weeks showing that an inverse-FEA predictor was held back by its inputs rather than the model, and the team stopped tuning and started fixing the measurements. I would rather report a ceiling than a number I cannot defend.

I build the product end too, since a model nobody can open is not finished. SeismicSoCal sends a real push alert when it detects a quake, BayRidership answers questions through a guarded SQL agent, and JobbyMail runs entirely on your own computer. Shipping forces the parts I would otherwise skip: latency, failure states, and what the thing does when the input is strange.

Toolkit

ML & deep learningPyTorch, scikit-learn, XGBoost, CNN / Transformer / GNN, PatchTST, neuralforecast
LLMs & retrievalRAG, LangChain, Ollama, Llama 3.1, embeddings, BM25 + vector hybrid search, rerankers, RAGAS
DataPandas, ETL pipelines, EDA, rolling CV, Matplotlib, Seaborn, Plotly, SQL
Data engineeringSpark (batch + Structured Streaming), Kafka, Iceberg, dbt, BigQuery, Airflow, Trino, Debezium, Terraform
BackendPython, FastAPI, Java / Spring Boot, PostgreSQL, Flyway, SQLite, Chroma
Frontend & desktopReact, TypeScript, Vite, Tailwind, Tauri (Rust shell), Capacitor (Android), hand-built SVG
Ops & MLOpsMLflow, Dagster, Evidently, GitHub Actions, Docker, systemd, Caddy, Oracle Cloud, FCM push
Statistics & causal inferencedifference-in-differences, event studies, A/B tests, McNemar, bootstrap CIs, conformal prediction
Testing & data qualitypytest, Playwright, Great Expectations, dbt tests, coverage, chronological validation