The work behind the work.
I build workforce intelligence: predictive turnover models, headcount and demographic forecasts, and the governed data layer they stand on. The modeling is rarely the hard part. Making a finding survive contact with the organization is, so that is where the effort goes.
Every project here says what it was for, what I made, when it shipped, and what changed because of it.
Why 92% accurate is not the number that matters
Say "everyone stays" and you score 90% without a model. Drag the slider to change how many people the model flags and watch which number actually moves.
Barely moves. The rose line is the do-nothing baseline.
Swings. This is the number a manager feels.
Accuracy barely moves. Precision is the number that decides whether a manager trusts the list, and it is the one nobody puts on the slide.
Illustrative population. The argument is developed in The Departures That Never Happened.
Selected work
Twelve projects. Each opens into the full story: where it started, what I made, when it shipped, and what changed. Employer and internal figures withheld by design.
Writing
Essays on measurement, prediction, and judgment at The People Data Lab on Medium. Ongoing since December 2025.
People Analytics Keeps Studying Employees and Forgetting to Study the Company
What gets preserved can quietly shape what gets explained.
How Many People Work Here? It Depends on Why You're Asking
What a headcount reconciliation taught me about Finance, People Analytics, and why making the numbers match isn't always the point.
Who Will Leave Is Not the Same Question as Who Can Be Kept
What building a turnover model taught me about the gap between prediction and intervention.
The Departures That Never Happened
A turnover model can be 92% accurate and still flag 40 of every 100 people who go on to stay. This is about those 40, and what they can teach you.
The Ethics Tax
The price of preventing your data from becoming a weapon.
Whose Story Gets Believed Once the Numbers Show Up?
How metrics pick whose reality counts.
The Work Before the Work
How to think through problems when the request itself is unclear. Introduces S.E.A.R.C.H.
The Dashboard Graveyard
A field guide to the analytics we build and never use.
Building a Turnover Model That Survives Contact With Reality
Why getting the model right is only 20% of the work.
People Analytics Has a Strategy Problem Disguised as a Skills Gap
What actually prevents analytics from influencing decisions.
New essays land on Medium first. Read everything at The People Data Lab ↗
About
Texas · Remote
People Analytics is applied behavioral science with heavy quantitative method underneath. Not a report desk.
More than a decade partnering with HR and business leaders on people data, across Big Four consulting, an energy company, and a global bank where I held a Vice President title in HR analytics and data governance. As a Senior People Analyst I have spearheaded most of what a People Analytics function is made of: the model, the forecast, the dashboard suite, the certified data layer, and the governance that lets all of it survive without me in the room.
The measurable part of the job is the smaller part. Whether a finding changes a decision depends on variable choices nobody sees, an audience chosen with care, a definition written down before the argument starts, and a willingness to not build something the data would technically allow.
Predictive modeling, scenario and time-series forecasting, anomaly detection, workforce planning, measurement and survey design, data governance. Currently building depth in survival analysis and causal inference.
Big Four consulting → energy → global bank, as Vice President in HR analytics and data governance → Senior People Analyst spearheading a People Analytics function.
R, Python, SQL, Snowflake, Tableau, Workday, Alteryx, Greenhouse, WorkTango. LLM-assisted workflows with reusable prompt frameworks for methodology and governance.
Global vice chair of a women's employee network, partnering with DEI on engagement and demographic analytics. Designed and taught an analyst-thinking curriculum for non-technical partners, built on
I/O psychology, causal inference, the economics of turnover, and the ethics of analytics that touch livelihoods.