← Back to portfolio

RESUME

Orhan Cenk Akcadoğan

Applied AI / Machine Learning Engineer

[email protected]LinkedInGitHub

Profile

Computer Science professional focused on applied AI, machine learning, and production-oriented AI systems. Experience spans NLP, computer vision, real-time ML, data analysis, APIs, mobile products, and cloud-backed applications, with end-to-end work from data preprocessing and model training to evaluation, deployment, and monitoring. Erasmus+ academic experience in Germany.

Experience

Data Scientist Intern

TFI TAB Food Investments · Istanbul, Türkiye

  • Applied Python-based data analysis and visualization workflows to explore business datasets and support data-driven reporting, including Power BI visualizations for analytical communication.

Data Scientist Intern

Istanbul Metropolitan Municipality · Istanbul, Türkiye

  • Worked with data preparation, exploratory analysis, and Python-based analytical workflows in a large public-sector environment.

Selected Projects

Real-Time Fraud Detection Platform

FastAPI, Kafka/Redpanda, PostgreSQL, XGBoost · Aug 2026 - Present

  • Engineered an end-to-end real-time fraud detection platform with streaming feature processing, low-latency scoring, PostgreSQL persistence, Docker Compose, explainability, drift monitoring, and an investigation dashboard.
  • Compared Logistic Regression, Random Forest, and XGBoost on 249,992 chronologically split transactions. The held-out champion achieved 0.900 precision and 0.3139 PR-AUC, with 19 automated tests passing.

Neural Singing Voice Platform

Python, PyTorch, FastAPI, Audio ML · Aug 2026 - Present

  • Engineered a modular Audio ML platform spanning audio validation, silence-aware segmentation, deterministic dataset manifests, pluggable model adapters, F0 analysis, conversion orchestration, gain-safe mixing, and content-addressed artifacts.
  • Implemented CUDA-aware device selection, durable SQLite job processing, FastAPI endpoints, a React/Vite interface, model registry, and repeatable pitch/audio evaluation.

Artist Decision Intelligence

Python, FastAPI, Multi-Signal Analytics · Aug 2026 - Present

  • Building a decision-intelligence platform that combines multi-provider chart, streaming, and audience signals into explainable local momentum and setlist recommendations.
  • Designed provenance-aware momentum analytics with provider-family caps, sparse and zero-signal handling, confidence and coverage controls, and D0/D7 validation to reduce noisy or misleading recommendations.

Skin Condition Classification & Detection

Python, Ultralytics YOLO, ONNX · Sep 2025 - Jan 2026

  • Built an end-to-end preprocessing pipeline for the SCIN dermatology dataset, trained and evaluated an Ultralytics YOLO model across seven categories, and exported the final model to ONNX for cross-platform inference.

Semantic Banking Assistant

Python, Sentence Transformers, Streamlit · Oct 2023 - Jun 2024

  • Built a Banking77 semantic assistant using all-mpnet-base-v2 embeddings, cosine-similarity ranking, confidence thresholds, top-k retrieval, cached model loading, and an interactive Streamlit interface.

Education

Mersin University

B.Sc. in Computer Science · Mersin, Türkiye

Hochschule Schmalkalden

Erasmus+ Exchange, Computer Science · Germany

Technical Skills

AI / ML: Sentence Transformers, NLP, Embeddings, Semantic Search, Ultralytics YOLO, scikit-learn, XGBoost, MLflow, SHAP, ONNX, PyTorch

Languages: Python, SQL, TypeScript, JavaScript, HTML/CSS

Data: pandas, NumPy, Data Preprocessing, Exploratory Data Analysis, Data Visualization

Backend / Cloud: FastAPI, REST APIs, Kafka/Redpanda, Supabase, PostgreSQL, AWS, Docker, Prometheus, Grafana

Frontend / Mobile: React, React Native, Expo, Next.js, Streamlit

Languages

Turkish: Native · English: Full Professional Proficiency