Software Engineer @ AI Collaborator — open to new roles
Full-stack engineer, four years remote across US and EU teams. I ship web platforms end to end — React and Next.js on the front, Node, Django and FastAPI behind, containers and cloud underneath — and I put LLM and machine-learning work on top of them. I came to software from energy systems engineering, and that is still how I think about a product: measure it, find the fault, fix the loss.
01 / About
I studied Energy Systems Engineering at Kocaeli University and spent my Erasmus+ placement at Ireland's Tyndall National Institute, building machine-learning models that spot faults in photovoltaic power plants before they cost anyone a kilowatt-hour. The research needed an interface, so I built one — and that is how I ended up a software engineer instead of an energy analyst.
Since then I have worked remotely for US-based teams, shipping customer-facing web products and the services behind them. The work I enjoy most sits where both halves meet: a real product that has to be fast and usable, with a model or a data pipeline doing something genuinely useful underneath.
02 / Skills
Interfaces that stay fast and usable once real data arrives — and that work for people who do not use a mouse.
Services and APIs across two ecosystems, so I can take the stack a product already has rather than insist on mine.
From the research side: classical ML on sensor data. From the product side: LLM features that have to be cheap and correct.
I deploy and operate what I build, which is usually the difference between a demo and a product.
03 / Experience
Features owned end to end: Next.js and TypeScript interfaces over Node, Express, Django and FastAPI services, deployed as containers on Docker, Kubernetes and OpenShift across AWS and GCP. Lately that has meant wiring LLM APIs, embeddings and retrieval into products people actually use.
Erasmus+ placement on fault detection in photovoltaic power systems: machine-learning models for monitoring and predictive maintenance, an automated data-processing and deployment workflow, and a full-stack application that let researchers see plant performance and faults without reading a CSV.
Headless-CMS eCommerce applications with GraphQL for efficient data querying, microservices and API integrations on cloud platforms, and MVPs taken from specification to launch — plus the unglamorous half: user administration, reporting and the internal dashboards that kept the platform running.
04 / Selected work
The photovoltaic work leads because it is the piece nobody else applying will have. The rest covers the range: a platform in production, a product of my own, and where I am heading.
Research · Real-time dashboard
Built at Tyndall National Institute: seven live views of a photovoltaic plant — array current and voltage, DC bus, three-phase current and voltage, frequency — reading from a Node and MongoDB service and redrawing as the day advances.
React · Chart.js · Tailwind · Node · MongoDB
See the project →Product · Mobile · Multi-agent
A personal investing companion I am building. Pick an asset and a set of background agents works out the short, medium and long-term technical picture, with a financial-literacy layer beside it. The analysis is meant to run whether or not anyone is watching.
React Native · Expo · Python agents · technical analysis
In developmentPlatform · Cloud infrastructure
The platform I work on day to day. I own a good part of its AWS side: Cognito and the identity flows around it, DNS and domain setup on Route 53, and the containerised services the product runs on. Under NDA, so the name stays off the page.
AWS · Cognito · Route 53 · Docker · Kubernetes
Ask me about itIn progress · Energy + AI
Where the two halves of my background go next: machine learning applied to renewable generation and energy-market data, starting from public datasets. The photovoltaic work at Tyndall was the first version of this question.
Python · forecasting · public energy datasets
Early stage05 / Education
Kocaeli University, Türkiye.
Tyndall National Institute / IERC, Cork, Ireland. Machine learning applied to photovoltaic systems.
06 / Certifications
Every title here links to the issuer's own verification page — Skilljar, Coursera, Credly, Udemy. Nothing to take on trust.
Full Stack Software Developer ↗
GEUBGVOC4T2Z · Coursera
The multi-course professional certificate covering front end, back end, containers and cloud deployment.
Forward Program ↗
Credly verified badge
Structured problem solving, communication and digital fluency — the parts of the job that are not code.
AI Summer Camp
Applied machine-learning programme, completed the same year as the Tyndall placement.
07 / Contact
I work remotely from Türkiye (UTC+3) and overlap comfortably with both European and US-East hours. Open to relocation for the right team.