Emre BULUT emrefbulut

About me

I'm an Electrical & Electronics Engineering student at Istanbul Medipol University, based in Istanbul, Türkiye. My focus is bringing AI, deep learning and LLMs into RF and microwave engineering

AI & deep learning — model training, fine-tuning, dataset engineering, LLMs
RF & microwave — SDR/IQ signal processing, SigMF pipelines, machine learning on RF
Hardware–software integration — edge deployment, embedded control, telemetry
Systems & infrastructure — virtualization, Linux administration, network topology

Featured Projects

IQForge

My first published Python package — pip install iqforge. Splitting SDR recordings at the window level lets neighbouring windows leak across train and test, inflating reported accuracy by up to 13.6 points. IQForge splits at the recording level instead, reads SigMF directly, balances metadata that would otherwise create hidden correlations, and fails loudly rather than degrading silently when stratification is impossible.

RoomGate‑AI

A smart access and occupancy system that turns detections into physical action. An edge-optimized YOLO26 detector with spatial region-of-interest filtering drives a relay-based lock, so the vision pipeline and the door stay in sync instead of drifting apart.

IoT Smart Energy Monitor

Three layers, one loop: an ESP32 + PZEM‑004T front end for raw AC measurement, a Python bridge computing active/apparent/reactive power and cumulative kWh, and a live dashboard on top. Anomalies raise alerts, events are written asynchronously to SQLite in WAL mode, and load shedding is guarded by hysteresis and a confirmation window.

VoltPilot

Sites planning EV charging investments rarely know in advance whether their grid connection can take the load. VoltPilot simulates it before anything is installed — transformer loading, battery dispatch, virtual grid signals, telemetry validation and generated engineering reports, so the decision rests on data instead of a guess.