I build systems that work where connectivity fails, compute is scarce, and latency is measured in milliseconds.
Full stack from PCB bring-up to cloud ingestion to production ML inference.
Domains: seismic & structural monitoring · edge computer vision · agentic LLM infrastructure
Languages
Python TypeScript Go C / C++ Node.js Bash SQL
AI / ML / Vision
PyTorch TensorFlow OpenCV YOLO v5–v11 ONNX TFLite HuggingFace LangChain
Embedded & Hardware
ESP32 (ESP-IDF) STM32 + FreeRTOS Raspberry Pi KiCad ADXL355 / IMU SPI / I2C / UART
IoT & Edge
MQTT over TLS RTSP AWS IoT Core OTA Offline-tolerant pipelines Edge GPU inference
Backend & Data
FastAPI Django REST WebSockets PostgreSQL Redis RabbitMQ InfluxDB MongoDB
DevOps & Observability
Docker Linux / systemd GitHub Actions GitLab CI Grafana Prometheus
Edge Computer Vision Real-time detection and classification pipelines with sub-100ms latency budgets. RTSP hardening for industrial conditions — lighting variance, motion blur, sensor noise. Multi-sensor fusion across camera, weight, barcode, and conveyor data streams.
Seismic & Structural Health Monitoring Signal processing and ML-based earthquake event detection. High-precision MEMS accelerometer arrays with device-to-cloud telemetry designed for reliability under field conditions. Vision Transformer approaches for P-wave magnitude estimation.
Embedded & IoT ESP32/STM32 firmware with watchdogs, fault recovery, and OTA. PCB prototyping from schematic through bring-up (KiCad). MQTT pipelines with offline tolerance and store-and-forward patterns for intermittent connectivity.
Agentic LLM Systems A2A Protocol runtime for multi-agent mesh architectures. Multi-LLM API gateways. MCP server tooling. Instrumenting agentic frameworks for production reliability.
| M.Sc. Civil & Structural Engineering | Ege University, İzmir — AI-Driven Embedded Systems for SHM & Predictive Maintenance |
| B.Sc. Civil & Structural Engineering | Ege University, İzmir |
| Exchange | METU, Ankara |
