Skills

The tools and methods I use to take a problem from question to a model people can trust, and the projects where each one is proven.

Programming languages

Python for modeling, pipelines, and production ML, SQL for everything data, R for statistical analysis, and JavaScript for interactive apps and tools.

  • Python
  • SQL
  • R
  • JavaScript

Deep learning & computer vision

Transfer learning and multi-scale fusion with CNNs and Vision Transformers, class-imbalance handling, and Grad-CAM to show what a network is looking at. Coursework spans YOLO detection and LSTM sequence models.

  • PyTorch
  • ResNet50
  • EfficientNet
  • Vision Transformer
  • Grad-CAM
  • YOLO
  • LSTM
Used inBreast cancer

Machine learning & ranking

Classification, recommendation, and learning to rank, from item-kNN baselines to LightGBM LambdaRank with multi-source candidate generation and leak-free temporal splits.

  • scikit-learn
  • LightGBM
  • LambdaRank
  • Feature engineering
  • Clustering
Used inRewear

Responsible AI & evaluation

Evaluation that keeps models honest: metrics weighted toward costly errors, bootstrap confidence intervals, beyond-accuracy metrics like coverage and novelty, and documented limitations.

  • Interpretability
  • Bootstrap CIs
  • Error-cost metrics
  • Model risk
Used inRewearBreast cancerFashion

LLM systems & generative AI

Production LLM pipelines and apps with validated structured outputs, offline fallbacks, and token budgets, plus RAG and Model Context Protocol integrations.

  • Claude API
  • RAG
  • Model Context Protocol
  • Prompt engineering
Used inDaily Companion

Statistics & forecasting

Time series, regression, and latent class models, with hypothesis testing and effect sizes that separate what is significant from what only looks suggestive.

  • statsmodels
  • SARIMA / SARIMAX
  • Regression
  • Latent class analysis
Used inRainfallFashion

Data engineering & MLOps

Scheduled ETL with quality gates, orchestration, containers, and infrastructure as code, with tests and coverage gates enforced in CI.

  • PySpark
  • Airflow
  • Docker
  • Terraform
  • GitHub Actions
  • pytest
Used inMile High SignalRewear

Databases & storage

Dimensional modeling and SQL across analytical and application workloads, from star schemas in Postgres to object storage on S3.

  • PostgreSQL
  • Star schemas
  • Amazon S3
  • Supabase
Used inMile High Signal

Visualization & apps

Charts that make a finding obvious, dashboards for non-technical stakeholders, and accessible front ends people can actually use.

  • Matplotlib
  • Seaborn
  • Power BI
  • React
  • Tailwind CSS
Used inFashionDaily Companion

Version control & documentation

Git and GitHub for every project, with READMEs and architecture decision records written so a teammate or interviewer can follow every trade-off.

  • Git
  • GitHub
  • CI/CD
  • Decision records
Used inMile High SignalRewear

Where each skill shows up

Skills are easy to list. This map shows the projects where each one was actually put to work.

RewearBreast cancerRainfallMile HighDaily CompanionFashionDenverSilicon Valley
Deep learning
Machine learning
Responsible AI
LLM systems
Statistics
Data engineering
Visualization

Have a problem worth modeling carefully?

I'm glad to talk about interpretable ML, LLM systems, or a dataset that deserves better.

Get in touchResume