4d ago · 13 min read · Imagine you're trying to describe a person to a friend. You could list ten thousand physical measurements—exact hair strand thickness, each fingernail length, the curvature of every knuckle—or you cou
Join discussion
May 27 · 9 min read · This is the fourth and final post in my Google Advanced Data Analytics Certificate series on the TikTok claims classification project. Part 1: data inspection. Part 2: EDA and visualization. Part 3: h
Join discussion
Apr 18 · 6 min read · A single regression model trained on NBA game logs predicts that Joel Embiid will play 11 minutes in a game where he's listed as OUT. The model has never seen a confident zero. Every row in the training data has some minutes played, because the stand...
Join discussionApr 15 · 26 min read · When building a machine learning model, it's tempting to train it on your data, check the score, and call it done. But that score is misleading — the model has already "seen" the data it's being evalu
Join discussion
Apr 12 · 14 min read · Where and how did it all start Imagine being handed a spreadsheet with 60,000 columns and being told — somewhere in here is the answer to classifying a cancer that kills 9 out of 10 patients within fi
Join discussion
Apr 7 · 12 min read · Feature engineering is where models are won and lost. It's also where ML teams waste enormous amounts of time arguing about tooling rather than building. The SQL vs. Python debate surfaces constantly, and most articles resolve it by quietly promoting...
Join discussion
Apr 7 · 13 min read · The bid request lands in your DSP in 100 milliseconds. You have maybe 10ms of that window to pull features, run your bidding model, and return a price. Meanwhile, the audience segment powering your decision was built from yesterday's batch job. That ...
Join discussion
Apr 7 · 12 min read · Credit decisions used to take days. A borrower submitted a paper application, a loan officer pulled a bureau report, an underwriter reviewed it overnight, and a letter arrived in the mail. Nobody questioned the timeline — it was just how lending work...
Join discussion
Apr 7 · 13 min read · Account takeover (ATO) is the fraud vector that breaks most traditional detection systems -- not because it is technically sophisticated, but because it uses entirely valid credentials. The attacker is not forging a card number or synthesizing a fake...
Join discussion
