The work, in long form.
A handful of projects, taken apart — a hardware build, a stretch of machine-learning research, and a run of text-mining and risk coursework. Each writeup is the honest version: what worked, what didn't, and why.
doorpi — a door that knows my face and waits for a peace sign.
Pi Zero + ESP32 + MediaPipe. A face-match unlock with a peace-sign verification step before the relay fires.
AmEx — when the classifier was most confident, suspect the label.
A research project with American Express. We used a BERT classifier's own confidence to hunt down the mislabeled tickets hiding in its "ground-truth" training data — and learned exactly where a second, fancier approach broke down.
Morning newsletters — telling the Times from the Brew was the easy part.
An independent-research project text-mining a year of two morning newsletters, scraped by hand. A classifier could tell which was which at about 89%. Predicting the date a newsletter went out — by topic, by token, by fine-tuned BERT — never worked at all.
The 2008 crisis — watching a bank's beta stop being diversifiable.
A risk-modeling project replicating Chaudhury (2014): rolling-window CAPM in R for Goldman Sachs and JP Morgan across 2006–2010. JP Morgan's beta climbed from 1.29 to 1.85 through the worst of it — and the real-estate funds we added on a hunch moved even more.
Oscar screenplays — spotting a winner was easier than naming the decade.
A text-mining project over 359 subtitle files — every Academy Award Best Original Screenplay nominee from 1940 to 2022. We expected the decades to separate cleanly and the winners to blur together. It was the other way around.