Data Scientist focused on supply chain & logistics analytics (forecasting, optimization, simulation) and financial risk analytics (climate risk, credit risk).
- 📍 Istanbul, Türkiye
- 💼 Data Scientist Associate @ PwC (Oct 2025 – Present)
- 📫 arda-orcun@hotmail.com
Links:
LinkedIn · GitHub · Kaggle
At PwC, I develop quantitative models to assess climate risk financial impacts and integrate them into PD/LGD frameworks; also support PD/LGD validation testing for audit engagements.
- Climate risk models → PD/LGD integration
- PD/LGD validation support
- Dynamic credit scoring pipeline for legal follow-up risk
- Automated reporting agent (LangChain + Azure OpenAI + SQLite)
- Meeting-scheduling agent (LangChain + Azure OpenAI + PostgreSQL)
- Multi-echelon distribution & factory-to-customer allocation (network-flow / MILP-style + heuristics)
- Cartonization & load optimization (bin packing / knapsack variants + heuristics)
- Forecasting + cashflow transformer pipelines
- ETL/CI-CD with Python, SageMaker, Redshift; global teams (MEA & Central Asia)
- Lightweight C++ GPU inference backend prototypes (latency & memory efficiency)
Note: Some work is internal/confidential; I can share patterns, architectures, and sanitized examples.
Turkey
- Forecasting transaction counts on a merchant basis — 2nd Place (Kaggle Leaderboard) [Feb 2024]
- Recommending new job listings based on applicants' previous applications — 2nd Place (Kaggle Leaderboard) [Feb 2024]
- TEKNOFEST Turkish Natural Language Processing Finalist [July–August 2024]
- Conference Presentation (2025): “Coreference and Entity-Aware Sentiment Analysis through Multi-Agent to Single-Agent Integration.”
Core: Python · SQL · C++
ML/NLP: PyTorch · TensorFlow · Transformers · Scikit-learn · XGBoost
Optimization/Simulation: MILP-style modeling · heuristics/meta-heuristics · scenario analysis
Cloud/MLOps: AWS (SageMaker, Redshift) · Docker · Git
Data: MongoDB
⚡ Fun fact: I'm the normal one.