I am a Computer Science student at Colorado State University, graduating May 2028, with a minor in Computer Engineering. My work spans machine learning, geospatial intelligence, embedded firmware, and full-stack development. I like problems that make you see the whole stack.
Born and raised in Turkey, I came to Fort Collins to study engineering. Before CSU I worked as a Junior Software Developer at Imecar Electronics in Antalya, shipping EV charging firmware and embedded systems. That work cemented my love for building things that touch the physical world.
Outside of tech I train MMA, grapple, and debate. They reward discipline and clear thinking under pressure, same as engineering.
Early education in an English-medium international curriculum that set the academic foundations for everything after.
Completed IGCSEs and A-Levels at one of England's oldest independent schools, with strong foundations in science and mathematics while living abroad for the first time.
Began my undergraduate degree at a Russell Group university, gaining early university-level CS experience before pursuing new opportunities in the US.
Transferred to CSU for Computer Science with a Computer Engineering minor. Since arriving I won a national geospatial AI hackathon, shipped industry firmware, and became an undergraduate TA. Graduating May 2028.
Professional roles and industry experience.
Support undergraduate Python coursework by writing course material and lab exercises, running lab sessions, holding office hours, and grading assignments and exams. I help students build strong programming foundations in an introductory CS curriculum.
Developed embedded firmware for EV charging infrastructure conforming to the DIN70121 and OCPP protocols. Built and maintained CAD models in SOLIDWORKS for hardware components, and acted as a technical interpreter and liaison for international partners, bridging language and specification gaps across teams.
Selected work across machine learning, embedded systems, and full-stack.
Won the 24-hour overnight CSU Geospatial AI Hackathon (USDA, NASA, Databricks, Esri, AWS) with a two-track ML system. Corn yield forecasting joins six independent sources into a 26,924-row Parquet across five states and twenty years, with per-forecast-date XGBoost regressors at +46.7% RMSE lift over baseline, calibrated uncertainty cones, and a four-tool Claude agent writing cited plain-English reports. Land-cover and GHG pairs a SegFormer-B1 model fine-tuned on LoveDA (val mIoU 0.5147) with an IPCC and EPA cited emissions calculator and a "what-if" land-use agent.
Production IoT systems shipped over nine months. A smart meeting-room display ran on an ESP32 with FreeRTOS, one core for FastLED, one for the web server, with SOLIDWORKS enclosures laser-cut from aluminium and a Node.js, PostgreSQL and JWT backend serving real-time room state.
Fine-tuned Qwen3-8B on five years of my own code into a local assistant that matches my style. A Python pipeline built a weighted dataset, then QLoRA via Unsloth trained it on an RTX 4070 Ti Super, quantized to GGUF and served through Ollama and Dockerized Open WebUI over Tailscale.
A dual dashboard for the USDA Plant Protection and Quarantine hackathon that finds high-risk fruit fly entry pathways into the US. An operational map fuses ArcGIS layers and trade data; a predictive dashboard runs SageMaker XGBoost with twelve deployed model artifacts and a live inference endpoint.
A cellular automaton engine with three variants: standard, toroidal wrap-around, and a three-state lifecycle. The board is a flat std::string for cache-friendly traversal, with an interactive pause and rollback buffer and a full getopt CLI plus test suite.
A cumulative recommendation engine in Java. It parses rating CSVs, then runs z-score normalisation, collaborative-filtering prediction, and K-means clustering, ranking in-cluster candidates by inverse distance for a top-twenty playlist across five CLI modes.
A real-time hand-tracking mouse controller with no inference-time model. It isolates motion with frame deltas in OpenCV, tracks offsets with NumPy, and maps hand position to the cursor through PyAutoGUI, with a planned TensorFlow gesture extension.
The stack I reach for, and what I am going deeper into.
Leadership, community, and a few things that keep me honest.
Have a project, a research opening, or just want to talk tech? My inbox is open.
