necatiincekara
30 verified skills
architecture
Design or analyze model architecture for this Ottoman QML study. Use for Turkish or English requests about V7-V10 ideas, bottlenecks, layer flow, parameter tradeoffs, or implementation planning.
data-ai
compare
Compare model versions, circuit variants, or training configurations in this repo. Produces evidence-backed tables and explains speed, stability, and accuracy tradeoffs.
testing
log-result
Log validated experimental results and update the study narrative without spreading stale claims.
testing
performance-debug
Diagnose training slowness, dead quantum signal, scheduler mistakes, and hybrid learning collapse in this repo.
development
reconcile-results
Reconcile contradictory metrics across docs, papers, notebooks, JSON logs, and checkpoints to determine the current factual status of the study.
testing
roadmap-sync
Turn benchmark truth, publication strategy, and compute constraints into a concrete next-phase study roadmap.
data-ai
setup-env
Set up or verify the development environment for local work or Google Colab training in this quantum ML repo.
development
status
Produce a current project status snapshot for this study, including code state, strongest results, stale docs, and next steps.
development
sync-colab
Manage code and result synchronization between the local repo and Google Colab without losing checkpoints or experiment metadata.
development
train
Orchestrate training for the baseline or enhanced hybrid models in this repo, including checkpoint handling and platform-aware execution choices.
testing
compare
Model, circuit veya egitim konfigurasyonlarini guncel sonuclara gore karsilastir. Tarihsel anlatidan degil, artefaktlardan ilerler.
development
experiment
Deney calistir ve sonuclari takip et. Ablation study, circuit karsilastirma, gradient analizi gibi deneyleri yonetir.
testing
gradient-check
Quantum layer gradient diagnostigi. Vanishing/exploding gradient, barren plateau tespiti yapar. V6 gradient collapse sorununu debug etmek icin kritik.
development
paper-sync
Makale, tez ve publication-facing metinleri mevcut repo kanitlariyla senkronize et.
testing
reconcile-results
JSON loglar, checkpointler, notebooklar ve dokumanlar arasindaki metrik celiskilerini cozerek calismanin bugunku gercegini bul.
testing
setup-env
Gelistirme ortamini platforma gore kur ve dogrula. M4 Mac ve Colab icin otomatik konfigürasyon.
development
status
Projenin guncel durumunu ozetle. Kod, deney, dokumantasyon ve yayin hazirligini bugunku gercege gore raporla.
development
train
Training yonetimi. Base, V7 trainable ve yerel ablation akislari arasinda dogru yolu sec ve calistir.
data-ai
architecture
Model mimarisi tasarla veya analiz et. Mevcut OCR-QML calismasinda trainable, non-trainable ve classical baseline yollarini guncel duruma gore ele al.
development
artifact-pack
Assemble advisor, submission, or sharing packets from current repo artifacts without introducing stale claims.
testing
benchmark-triage
Decide the next benchmark or training action using current artifacts, compute budget, platform constraints, and paper impact.
testing
codex-audit
Audit official Codex capabilities against this repo's current integration so you can identify underused or stale Codex layers safely.
development
experiment
Plan or run an experiment for this study, including ablations, circuit comparisons, gradient diagnostics, and result logging.
testing
log-result
Deney sonucunu guncel artefaktlarla uyumlu sekilde kaydet. Stale claim uretmeden docs ve result loglarini gunceller.
documentation
paper-sync
Synchronize paper and thesis writing with the actual experimental evidence in this repo. Use before drafting, revising, or finishing the paper.
research
performance-debug
Yavaslik, scheduler hatasi, dead quantum signal veya hybrid egitim cokusu gibi repo-ozel performans sorunlarini analiz et.
development
review-circuit
Review a quantum circuit or hybrid QML block with repository-aware best practices for expressivity, gradient flow, and training stability.
testing
review-circuit
Quantum circuit implementasyonunu uzman gozuyle incele. Barren plateau, expressivity, gradient flow, gate optimization analizi. quantum-ml-reviewer agent'ini kullanir.
development
sync-colab
Mac ve Google Colab arasinda kod ve sonuc senkronizasyonu. GitHub push/pull, Colab ortam hazırligi, checkpoint transferi.
development
gradient-check
Diagnose vanishing or exploding gradients in the hybrid quantum-classical training path. Use especially for V6 collapse, trainable quantum instability, or AMP-related failures.
testing