A bit about me
I’m a data scientist and software engineer who works where statistics meets production engineering. For the past several years I’ve led R&D work at a medical technology company, building machine learning that ends up in the operating room: imaging pipelines, sensor fusion, and LLM-based tools that help clinicians find what they need.
I started out in aerospace, writing software for a satellite and for sensor pipelines where a wrong number is expensive. That shaped how I work today. A model only matters if the software around it is reliable, easy to maintain and measured against real outcomes.
Here I write about machine learning, medical imaging, signal processing and the math underneath them. I speak Turkish, English and German. Away from the keyboard I follow aviation, watch clouds, collect fountain pens and watch a lot of films.
Experience
- 2020 – nowSenior R&D Engineer, Team Lead · Medical technology companyLeading a team of five that builds production ML for surgical and imaging products, from RAG-based LLM tools to computer vision pipelines and statistical models. I own architecture decisions and see projects through to release.
- 2019 – 2020Senior Software Engineer, R&D · Medical technology companyBuilt ML software that combines MR images with sensor data captured during surgery, giving surgeons real-time guidance in the operating room.
- 2017 – 2019R&D Software Engineer · Aerospace companyDeveloped orbital systems software for a national communications satellite in Fortran 95, C, C++ and Python. The national research council reviewed and recognised the work.
- 2016 – 2017Software Systems Engineer · Satellite operatorWorked on load balancing for a satellite internet service and led efforts to improve its wireless delivery, which brought me onto the satellite program.
- 2012 – 2014InternshipsData science, engineering and avionics internships at a national statistics institute, a global technology company and an aerospace company.
Consulting Experience
- 2024 – 2025Energy cost optimisation platform · IrelandLed the architecture and analytics of a utility-bill and tariff comparison platform. Built its solar billing system, which pulls together many data sources and large volumes of data for accurate billing, and a tool that estimates a home's rooftop solar potential, investment cost and payback period. Added LLM agents that automate sign-up forms and switch customers to cheaper plans, plus a RAG-based WhatsApp assistant for customer support. The work cut manual billing time considerably and raised solar tariff adoption.
- 2021 – 2022Textile and embroidery manufacturer · TurkeyWrote software that lets a sequin embroidery machine treat sequins like pixels when stitching an image onto fabric, and helped integrate it into the existing production line. Also advised on making production processes more data-driven and efficient.
Education
- 2014 – 2016M.Sc. Computer Engineering · Technische Universität BerlinMachine learning and data science. Thesis: spotting faults early in high-frequency, noisy sensor data with deep learning models, for parts like aircraft engines and turbines.
- 2009 – 2014B.Sc. Computer Engineering · Middle East Technical UniversityGraduated with high honours. Graduation project on machine-learning detection of DDoS attacks in WiMAX networks, first place at the METU Demo Day 2014.
Highlights
- 2026Patent pendingA detrending-based method that builds local phase and magnetic susceptibility maps from complex MR data to help tell tissues apart.
- 2022Oracle Cloud Infrastructure certificationsOCI Architect Professional and OCI Solution Architect Associate.
Tools I use
- Languages
- PythonCC++JavaFortranSQL
- ML & AI
- PyTorchscikit-learnTensorFlowTransformersLLM / RAGLangChainComputer visionTime seriesNLP
- Data
- PostgreSQLpgvectorPineconeSnowflakeAirflowKafkaPower BI
- Cloud & ops
- AWSGoogle CloudDockerKubernetesMLflowCI/CDPrometheusGrafana