Stack

No skill percentages here — a bar chart can't prove anything. Instead, a map: every edge connects a tool to a project where it was actually used.

fig. 2 — every edge is a verified use, not a skill bar. Signals flow from tool to project. Hover a node; click a project to open its case study. Dashed = still learning. Scroll to zoom, drag to pan.

In daily use

Pythonused in: every project on this site
Pandasused in: thesis EDA · Telco churn · auto sales
NumPyused in: thesis pipeline · CV preprocessing
Scikit-learnused in: thesis models · Telco churn
PostgreSQLused in: PUSULA web app (Neon)
Power BIused in: sales analytics dashboards
OR-Tools CP-SATused in: shift solver in production · PUSULA
FastAPIused in: shift-solver API (/solve, /solve/area-based)
Next.jsused in: this site · skdesign.online
Reactused in: PUSULA web app (React 19)
Git / GitHubused in: all repos
MLflowused in: Telco churn run tracking
TypeScriptused in: this site · Budgeify · Prexamy web · skdesign
Flutterused in: Prexamy mobile app (BLoC, v1.0.3)
NestJSused in: Prexamy API — AI orchestration layer

Currently learning

PyTorchlearning — HRV/ECG work at Bluesense
TensorFlowlearning — coursework
OptunaTeknofest recsys hyperparameter search
C#CrmApp desktop CRM — university coursework (NTP-2)
SQLiteCrmApp embedded DB — schema, FKs, seeded CRUD