Machine Learning Researcher — interpretable ML, structured NLP, and financial time-series. Computer Engineering undergraduate at METU applying transparent, reproducible methods to monetary-policy decisions, central-bank communication, and financial markets, including public research infrastructure built from thousands of central-bank documents and cross-country models spanning dozens of economies.
Through the Hidden States: Synthetic Latent Supervision for Sparse-Label Financial Prediction — Accepted, ICBFS 2026. Latent-state transfer framework cutting forecast error 6-7% in data-scarce financial markets. Manuscript · Code
No One Gets an A: Distributional Bias in LLM Text Scoring — Under review, ACL ARR. Audits 11 LLMs across ~99,000 evaluations for score compression. Manuscript · Code
A Gaussian Recency Score for Central-Bank Policy-Rate Timing and Direction — Under review, Journal of Forecasting. Interpretable policy-rate timing/direction model across 48 economies. ResearchGate profile
Central bank communication graphs for Bank of England and Federal Reserve documents, 1998-2026 — Under review, Scientific Data. CC BY 4.0 temporal graph dataset from 2,872 documents. ResearchGate profile
Attributing Central-Bank Communication: Content Channels and Tail Repricing of UK Rate Expectations — Working paper. Typed-graph attribution linking BoE communication to rate-expectation repricing. ResearchGate profile