Self-taught projects · Writing
Before the career, there was a lot of teaching myself in the open — building projects and writing them up so the fundamentals would click for other people learning ML.
The projects
Two that I still like: an unsupervised LDA topic-modelling project on Wikipedia forum comments that isolated toxic comments (insults, slurs) into their own cluster and labeled posts into categories; and a clustering study (KMeans and agglomerative, in scikit-learn) on a tech-industry mental-health survey, practicing categorical encoding, visualization, and cluster-validity assessment.
The writing
I wrote two pieces that found a real audience: “How I analyzed a million Medium.com articles” on freeCodeCamp, and “How to Learn Data Science If You're Broke” on Towards Data Science — a guide to teaching yourself the field on free and low-cost resources. More on Medium.
What it demonstrates
The self-directed learning habit that never really switched off — and an instinct for explaining technical ideas simply, which turns out to matter as much as building them.