Machine Learning Engineer
Muhammed Mustafa KAPICI
I am a Computer & Software Engineer specializing in Machine Learning, Deep Learning, Computer Vision, and Large Language Models, with hands-on experience building scalable, production-ready AI systems end-to-end. I have worked across space technologies, mobile development, and microservice-based AI platforms, delivering solutions that combine advanced model architectures with modern engineering best practices.
Experience
Software Engineering Intern (Internship)
Entegre Yazılım
- Developed a fully functional medication tracking mobile app using React Native & Expo, designed for elderly-friendly usability and offline reliability.
- Integrated Google Vision OCR for automatic prescription label scanning and built an AI-powered scheduling module using DeepSeek LLM for adaptive reminder generation.
- Implemented local notification pipelines, modular state management, and prepared the application for multi-user and wearable device integration.
Computer Vision Intern (Internship)
TUBITAK SPACE Technologies Research Institute
- Led the development of an astronomical object detection pipeline specialized in Point Spread Function (PSF) analysis for enhancing star and galaxy identification in FITS images.
- Built a custom Faster R-CNN model with PSF-aware layers and integrated Hough-based feature extraction.
- Improved detection precision significantly by leveraging distributed training (PyTorch DDP).
- Implemented SEP and Photutils tools for source extraction and used Weights & Biases for advanced experiment tracking and metric visualization.
Software Engineering Intern (Internship)
Kardelen Yazılım
- Contributed to the development of enterprise application modules using C# .NET, Entity Framework, and MSSQL Server.
- Enhanced database operations, optimized query performance, and collaborated with senior engineers to improve backend reliability.
- Supported UI and business logic components, ensuring consistency across the software ecosystem.
Projects
CreditCard-clustering
Developed a full unsupervised learning pipeline for credit-card customer segmentation. Includes preprocessing (imputation, winsorization, scaling), engineered behavioral features, dimensionality reduction (PCA, UMAP), and clustering with K-Means, GMM and HDBSCAN. Integrated model evaluation (silhouette, DB index, ARI) and built a Streamlit dashboard for interactive exploration.
IRIS-classification-pytorch-mlflow-MLOps
Built a production-style PyTorch training framework with MLflow experiment tracking and Model Registry. Added a FastAPI inference service with Prometheus metrics, Dockerized workflows, Makefile automation and dual frontends (Streamlit + Gradio). Demonstrates a complete MLOps-ready lifecycle from training to deployment.
Boston-House-prediction-XGBOOST-MLFLOW
Architected a high-performance regression pipeline using XGBoost for real-world price prediction, achieving a 0.94 R2 score and a low MAE of 1.88. Features preprocessing, domain-specific feature engineering, hyperparameter tuning, model explainability, versioned model management and deployable artifacts.
Gemini-Langgraph (RAG System)
Developed a full-stack Retrieval-Augmented Generation system using LangGraph, Google Gemini and ChromaDB. Includes a modular document ingestion pipeline, embedding + retrieval workflow, streaming-enabled FastAPI backend, a React/Vite frontend, and Docker-Compose orchestration. Demonstrates modern LLM engineering and production-grade RAG design.
UCI-heart-disease-predictor (MLP / XGBoost / Logistic Regression)
Created a multi-model classification pipeline for clinical heart-disease prediction. Covers preprocessing, feature engineering, stratified cross-validation, hyperparameter tuning and probability calibration. Combines neural networks (MLP) and tree-based models (XGBoost) to analyze structured medical data.
crypto-forecaster-LSTM-GRU-Transformer
Built a high-frequency time-series forecasting engine for crypto markets using LSTM, GRU and Transformer architectures. Includes automated market data ingestion, technical-indicator generation, volatility analysis, S2S modeling, hyperparameter optimization and full backtesting simulations tailored for financial prediction tasks.
EuroSAT-Classification-Swin-ConvNeXt
Designed an end-to-end satellite image classifier using state-of-the-art architectures (Swin Transformer, ConvNeXt). Features mixed precision, warm-up cosine LR scheduling, weighted sampling/Focal Loss for imbalance, patch-based inference, TTA, and a deployable Gradio UI. Demonstrates advanced CV techniques on remote sensing data.
Astro-psf-detection-resnet
Engineered a deep-learning pipeline for astronomical PSF detection using a customized ResNet backbone. Workflow includes simulation-driven augmentation (blur, atmospheric distortion), domain-adapted preprocessing, fine-tuning, and morphological post-processing. Shows capability in bridging astrophysics and convolutional networks.
mkyz (Python ML Library)
Created a lightweight, pip-installable library that simplifies ML workflows: preprocessing utilities, classification/regression/clustering models, automated evaluation metrics, and integrated plotting. Designed for rapid prototyping and educational or experimental ML tasks.
Skills
Education
Bachelor’s in Computer Engineering & Software Engineering
Toros University — GPA 3.14/4.0