Izmir, Türkiye / Ege University, Mathematics (CS-weighted), 2021–2026
Ahmet BaranBozkurt
Data scientist in the making — mathematics graduate who builds real systems, now looking for a production-scale role. Every number on this site links to its source — if it isn't verified, it isn't here.
fig. 0 — pusula / compass
§ 01 — Numbers
- GPA out of 4.00 — 4th in department
- 3.44
- evidencetranscript · Ege University
- simulated conversion lift (test set)
- +26.5%
- evidencethesis, Table 5, p. 40
- shift-solver overlap vs 10 real charts
- 63.8%
- evidenceparsers/backtest.py · README
- images processed, Teknofest CV
- 698
- evidenceteknofest-traffic-cv dataset
§ 02
Selected Work
the complete index
11 / 11
§ 03
About
Ahmet Baran Bozkurt
Izmir, Türkiye
Ege University, Mathematics (CS-weighted), 2021–2026
Data scientist in the making — mathematics graduate who builds real systems, now looking for a production-scale role.
None of that happened in a quiet study. I put myself through university working part-time at ZARA (Inditex) and moved to full-time the moment I graduated — running the shop floor and the coursework at the same time for years. In parallel I led a team at Teknofest, shipped side projects, and took a volunteer internship at Bluesense AI. Doing the degree, the retail job and the building all at once is where I learned to adapt fast, absorb a new domain quickly, and belong to a team rather than just occupy a seat on it.
§ 04
Writing
- 2026-07-05Accurate Offline, Useless In-App: The Tradeoff Nobody Grades You OnDuring a volunteer internship at Bluesense AI I integrated Hugging Face face-analysis models into a real product and built HRV/ECG signal pipelines. The lesson that stuck wasn't about accuracy — it was that a model's benchmark score and its in-app usefulness are two different numbers, and only one of them ships.
- 2026-07-04Scheduling a Retail Floor with CP-SAT: An Honest Shift-Optimization ArchitectureHow I modeled hourly staff placement in a ZARA store as a constraint-programming problem with Google OR-Tools CP-SAT, served it over FastAPI, validated it against ten periods of real human decisions — and how the solver then outgrew the hackathon and became the scheduler my section actually uses every day.
- 2026-07-04When Your Model Scores a Perfect 1.0, Don't Celebrate — InvestigateMy first conversion model scored a perfect ROC-AUC of 1.0. It wasn't skill — it was target leakage. This is how a six-scenario ablation found the culprit, and why the honest 0.71 turned out to be worth more than the fake 1.0.