// Data Analysis, ML, APIs, and Practical AI Systems

William Huddleston

Bachelor of Science in Information

Major: Information Analysis

Minor: Human-Centered Artificial Intelligence

GPA · 3.95Expected Graduation: May 2028
Seattle, WA / Ann Arbor, MI

About

I’m a data-focused student at the University of Michigan School of Information, specializing in Information Analysis with a minor in Human-Centered Artificial Intelligence. I build pipelines, analyze datasets, and ship AI-powered tools that solve real problems.

This past summer I worked on the Enterprise AI team at McKinstry, a national construction and energy services firm, building internal tools on the Claude API and deploying them to Azure. Before that I shipped a production voice assistant for a specialty retailer and an EDI monitoring system for an automotive supplier.

I also run Huddleston AI, a small consulting practice where I find the workflow a business wastes the most time on and automate it. Across all of it I care about clean data, clear communication, and systems that hold up once real people depend on them.

Outside of data science, I’m a big football fan, especially college football and the NFL. You can always catch me watching a Seahawks game! Go hawks and go blue!

Education

University of Michigan School of Information

Relevant Coursework

Multivariable CalculusData-Oriented ProgrammingStatisticsData ManipulationSQL & DatabasesUnderstanding AI

Experience

Where I've applied data analysis and AI.

McKinstry

AI Intern, Enterprise AI Team

May 2026 – Aug 2026 · Seattle, WA

  • Built an AI-powered change order narrative and estimation tool for construction project managers using Python and the Anthropic Claude API, taking a process that ran about two days down to under an hour. Piloted with 5 project managers.
  • Deployed internal AI applications to Microsoft Azure on Function Apps with Cosmos DB and PostgreSQL backends, establishing the deployment patterns for the company's enterprise AI initiative.
  • Ran McKinstry's first internal AI hackathon with 30 participants, and trained employees across construction, engineering, and IT on using Claude Code in the terminal and in VS Code.
  • Mapped company-wide SaaS data systems including Salesforce, Viewpoint, and Pacific Timesheets into a Microsoft Fabric ingestion plan.

Michigan Data Science Team

Data Analyst

Jan 2026 – May 2026 · Ann Arbor, MI

  • Completed the team's data science onboarding track, then joined the Winter 2026 clinical modeling project.
  • Worked a nine-week research cycle on a real patient dataset, from exploratory analysis and statistical testing through to a final evaluated model.

Michigan AI Business Group

Project Manager / Engineering Analyst

Sept 2025 – Apr 2026 · Ann Arbor, MI

Clients: Sigma International, Cherry Republic

  • Led an EDI release monitoring system for an automotive supplier, comparing incoming order schedules against committed quantities to catch late changes inside the lead-time window. Projected to cut expedited air freight costs by 20 to 30% annually.
  • Built Tableau and Power BI dashboards over thousands of EDI transactions to surface change frequency, cost impact, and operational risk.
  • Developed an automated Python notification system targeting sub-5-minute alert delivery on high-risk order changes.

GamePoint Capital

Data Analyst

May 2025 – Sept 2025 · Remote

  • Manually collected, cleaned, and structured NCAA sports data from multiple sources to support predictive analysis.
  • Analyzed performance trends across 350+ teams and 15+ sports using historical metrics and standardized datasets.

Terra Health Essentials

Research Intern

June 2023 – Aug 2023 · Seattle, WA

  • Analyzed and synthesized findings from 15+ peer-reviewed clinical studies, extracting quantitative and qualitative evidence to support product research.
  • Structured research outputs into comparative summaries used to inform internal decision-making across 10+ supplement products.

Skills

Tools and technologies I work with.

Languages

PythonSQL (PostgreSQL, MySQL, SQLite, DuckDB)JavaScript

Data & ML

PandasPolarsNumPySciPyscikit-learnLightGBMPyTorchOptunaSHAPExploratory Data AnalysisFeature EngineeringHypothesis Testing

AI Engineering

Anthropic Claude APIClaude CodeTwilioElevenLabsNLTK

Cloud & Deployment

Microsoft Azure (Function Apps)Cosmos DBMicrosoft FabricVercelGit / GitHub

Data Engineering

REST APIsETL PipelinesData ScrapingSchema DesignIndexing

Web & Visualization

ReactNext.jsTableauPower BIMatplotlibSeabornExcel

Projects

Featured work and case studies.

Heart Failure Survival Analysis cover

Heart Failure Survival Analysis

Data Scientist

Benchmarked six classical machine learning models against three deep learning architectures on a real clinical dataset (n=299, 12 features) to predict heart failure survival, then used interpretability analysis to work out which clinical variables actually drive mortality risk.

  • Tuned with Optuna Bayesian search to reach 0.80 ROC-AUC
  • Ran 30+ statistical tests with Lasso, Elastic Net, and MRMR feature selection
  • SHAP analysis isolated the top 5 mortality predictors, led by ejection fraction and serum creatinine
Pythonscikit-learnLightGBMPyTorchOptunaSHAP
Cherry Republic Voice Assistant cover

Cherry Republic Voice Assistant

AI Developer

Designed and deployed an automated voice assistant for Cherry Republic that handles 800+ missed customer calls per month, using Twilio and ElevenLabs for natural voice interactions.

  • Automates intent capture, order questions, and 15+ FAQ workflows
  • Rule-based call flows and structured logging across 20+ call scenarios
  • 100% after-hours coverage, with monitoring to support iteration
TwilioElevenLabsPythonAI
College Football Weather Analytics cover

College Football Weather Analytics

Data Analyst

Built a Python data pipeline integrating four APIs and a normalized SQLite database to analyze 125+ NCAA games, investigating the relationship between weather conditions and game scoring.

  • Integrated 4 APIs into a normalized SQLite database
  • Pearson r = 0.42 temperature-scoring correlation via SciPy
  • Created bar charts, box plots, and heatmaps for clear communication
PythonSQLitePandasSciPyMatplotlib

Get in Touch

Interested in working together? I'd love to hear from you.

Send an Email