PORTFOLIO / 2026

I build models that turn noisy data into decisions.

Data scientist working across applied mathematics and statistics — research, open-source tooling, and production systems that hold up under scrutiny.

Read the posts See the work →
3.3σ 3.6σ 5.8σ feature 2 feature 1 inlier anomaly fig — robust outlier detection 5 / 64 flagged
PROJECTS · selected builds All projects →
Live
Experimentation · Tool · Bayesian methods

Positive Predictive Values

A small webapp that let's you play with base rates, true positive, and false positive results to see how the math of Bayes Theorem corrects for the Base Rate Fallacy and our broken intuition around conditional probabilities.

PythonMarimo
Open project ↗
No matching items
POSTS All posts →
No matching items
PORTFOLIO All work →
Portfolio project · 2026 Quantified Risk
Modeling the Customer Lifecycle: One Dataset, Five Questions
No matching items