ML & Data Science
3 results Clear
AnalysisFrom Notebook to Production · Analysis98% Accurate and Useless: Six Imbalance Fixes, Measured on a 2% Positive Class
Six models on one 2% positive dataset: the baseline caught 1 of 59 positives, class weights caught 45 at 6% precision, and a balanced random forest predicted zero positives despite the best AUC. Threshold tuning was the only fix that raised F1.
Ibrahim Denis Fofanah·Nov 12, 2026·11 min
AnalysisModel Evaluation · AnalysisSame Model, Ten Seeds, a 3.5-Point Swing: Why One Split Is Not a Result
Same data, same code, ten split seeds. Logistic regression ranged from 95.8% to 99.3%, and the "winning" model flipped on two seeds. A single split is not a result.
Ibrahim Denis Fofanah·Nov 2, 2026·8 min
AnalysisData Engineering · PerformanceOpenAI Got 6× Better CPU Efficiency With Rust. The Lesson Is Why They Waited.
OpenAI’s storage service once pushed more than 20 million requests per second through Python. A Rust rewrite later delivered 6× better CPU efficiency and 15× better memory efficiency. The useful engineering lesson is not “rewrite everything in Rust.” It is why OpenAI knowingly carried the slower system first.
Ibrahim Denis Fofanah·Sep 25, 2026·16 min