Writing
Technical articles and case studies on ML, web development, and software engineering
CoreMetric: ML-Powered System Monitoring on macOS
Building a privacy-first system monitor that uses neural networks and the Apple Neural Engine to detect anomalies. Combines PyTorch, Metal Performance Shaders, CoreML, and SwiftUI to deliver intelligent monitoring with <1% CPU overhead.
My Machine Learning Journey: From NumPy to Production PyTorch
A personal roadmap through machine learning: Python foundations, NumPy, Pandas, scikit-learn, PyTorch with Apple Silicon MPS, deep learning architectures, MLOps, and production deployment.
GitSimulator: Production-Grade Git Simulation Engine
Deep technical dive into GitSimulator (3rd Place, GitKon 2025 Game Jam): 135+ tests, automated CI/CD, interactive TUI, plugin architecture, and advanced conflict prediction.
LectureLens: AI Study Planner with Next.js + Supabase
Open-source study companion: ingest course materials to generate summaries, flashcards, Q&A, and schedules — built with React, Next.js, and Supabase.
Next.js Performance Optimization: 58 → 100 Lighthouse Score
Complete guide to achieving 100/100 Lighthouse score in Next.js. Real portfolio case study: WebP optimization, Three.js deferring, code splitting, accessibility, and security headers.
CDN Optimization with AWS CloudFront + S3: A Real-World Case Study
How I shipped a fast, secure, and low-cost static site on S3 + CloudFront: OAC, cache policies, versioning, and concrete performance gains.
PostgreSQL Query Optimization: 40% Performance Boost in Production
Real-world PostgreSQL optimization from handling 50,000+ daily sensor readings: indexing strategies, query planning, partitioning, materialized views, and monitoring.
Case Study: CoreMetric — Neural Anomaly Detection on macOS
Privacy-first system monitor using Reconstruction Autoencoders on Apple Neural Engine. Complete ML pipeline from PyTorch training with MPS to CoreML inference in SwiftUI with <1% CPU overhead.
Case Study: GitSimulator — Flight Simulator for Git Commands
Production-ready Python CLI with 135+ tests, automated CI/CD, interactive TUI, and plugin architecture. Safely simulates rebase, merge, reset & cherry-pick with advanced conflict prediction.
Case Study: DigiTwin Database System — 50K+/day, 40% Faster Queries
High-performance database for structural monitoring: 50,000+ daily readings, 40% faster queries via partitioning and materialized views, and a robust C++ ingestion pipeline.
Case Study: LectureLens — AI Study Planner (Open Source)
(WIP) Open-source AI study planner: ingest PDFs & notes, generate summaries, flashcards, interactive Q&A, and structured schedules (Next.js + Supabase).
Case Study: egekaya.dev — Next.js 16 + Tailwind v4
A fast, accessible, and maintainable portfolio built with Next.js 16, React 19, TypeScript 5, and Tailwind CSS v4. Includes Three.js background, retro mode, and robust writing UX.
Case Study: Parma Calcio Internal Platform
Delivering a secure, high-performance internal platform for Parma Calcio 1913: automated workflows, edge-cached docs and videos via CloudFront + S3, and measurable time savings.
Case Study: Discord Bot Automation for PoliTo Organization
Automating role assignment for 200+ members, reducing manual management time by 95% using Discord.js and Supabase.