Three applications taken from model to deployment and still running, the modelling work that lives as notebooks, and — so the selection is legible — what this page leaves out and why.
Fig. 1NVDA — the price lenses and the filings lenses disagree.
A quantitative equity research desk that runs four models reading four different datasets — order flow, price trend, intrinsic value and accounting quality — then reports where they agree, where they contradict each other, and what none of them can see. Covers US and Indonesian (IDX) listings.
Fig. 243 programmes, filtered by what actually disqualifies you.
A zero-cost dashboard for finding, filtering and tracking fully-funded international scholarships. Programme details live in structured JSONB columns rather than prose, so questions a wall of text cannot answer become filters — and every extracted value is verified against a verbatim quote from the funder's own page before it is allowed into the database.
Fig. 3Progress, with every unmeasured figure saying so.
A PWA that teaches English and Japanese to Indonesian speakers — no account, no paywall, no ads, no engagement mechanics, and zero recurring cost. A learner on a cheap Android phone taps the icon and is answering a useful question in 108 ms, fully offline, because a memory model predicts they are about to forget it.
Work held as notebooks rather than products. Listed with its method and data, which is what makes an analysis judgeable without opening it.
A list of only the good things tells you nothing about how it was selected, so these are named rather than quietly omitted.
Finished applied-ML work that has never been pushed to a public repository. Left out because work a reviewer cannot open is an assertion rather than evidence.
Repositories that exist and are readable, but are not finished work. They are not listed as projects, and no README has been written to make them look finished.
Work that exists and is complete, but is not mine to publish.