About
I'm an AI engineer who also writes a lot of software. I take a model that works on my laptop and turn it into something other people can use.
Lately that has meant a service that screens posts for abuse without slowing anyone down, AI agents that boil a 100-page report down to two pages, camera models that spot theft, fire and crowding, and a detector for faked faces.
Before that I spent 17 months at Levant Yarn in Saudi Arabia, running ad experiments and keeping the IT side working. It taught me to check the numbers before arguing about them, and that people depend on what I ship.
Now
Building a RAG assistant over my own notes and reports. It shows which passages each answer came from, and I'm scoring how often those answers are right.
Education
- B.Sc. Artificial Intelligence EngineeringBahçeşehir University, Istanbul2023 – 2027
Roles
- Trained camera models to spot theft, fire and crowdsComputer vision & AI engineering intern · GateworxJul – Aug 2026
- Ran ad experiments, studied the sales funnel and kept IT runningGrowth data analyst & IT specialist · Levant Yarn, Saudi ArabiaApr 2024 – Aug 2025
- Worked out who engages with campus events, and when to promote themData scientist · NEST club, BAU2024 – 2027, until I graduate
- Founded and ran a regular reading and discussion seriesFounder · The Book Club, BAU2025 – 2026
Tools
AI / LLM engineering
- LangGraph
- LangChain
- RAG
- OpenAI & Anthropic APIs
- Hugging Face
- pgvector · Qdrant
- Embeddings & reranking
- LLM evals (Ragas, LangSmith)
- Tool calling · MCP
- Pydantic
ML / deep learning
- PyTorch
- TensorFlow
- Ultralytics YOLO
- OpenCV
- Transfer learning
- LoRA fine-tuning
- scikit-learn
- XGBoost
- MLflow
- ONNX
Data science
- Python
- SQL
- pandas
- NumPy
- Statistics
- A/B testing
- Cohort & funnel analysis
- Jupyter
- Power BI
Full-stack & infrastructure
- TypeScript
- Next.js · React
- FastAPI
- Django
- PostgreSQL
- MongoDB
- Redis
- Celery · RabbitMQ
- Docker
- Kubernetes
- AWS · GCP
- GitHub Actions
- pytest