CS student at Colorado State University and undergraduate researcher in CSU's Quantum Research Lab. I build AI agents, machine-learning and computer-vision systems, web products and connected hardware, end to end.

🌐 gedik.tech · 💼 LinkedIn · 📍 Fort Collins, CO

• Terra, 1st place at the CSU × USDA × NASA Geospatial AI Hackathon 2026. County-level corn yield forecasting from NASA satellite data: six fused data sources, per-date XGBoost (+46.7% RMSE over baseline), calibrated uncertainty, SHAP leakage checks, SegFormer land-cover segmentation, and an LLM agent that cites numbers instead of inventing them.
• Qupacabrathon 2026, 1st place and Largest Certified Cycle at CSU's quantum hackathon. When the hardware API went down, we rebuilt the device error from IQM Emerald's calibration data (a noise model across all 86 qubit pairs) and sized our bet on it.

• TuneSummary: free listening stats from your Spotify data export, live at gedik.tech/tunesummary. FastAPI + SQLite, in-browser export parsing, Last.fm live sync, and a canvas recap story with song previews.

• AI operations team: 14 agents on 3 isolated gateways, reachable over Telegram, WhatsApp and Discord, with approval buttons, a watchdog, an auditor agent and nightly encrypted backups.
• Private code LLM: Qwen3-8B fine-tuned on my own code with QLoRA (Unsloth), quantized to GGUF and served locally through Ollama.
• Computer vision for a client: face and physique analysis from ordinary photos (ONNX, OpenCV), plus identity-consistent image generation (ComfyUI, LoRA, ControlNet).
• Imecar: 9 months of industry IoT and CAD work, including EV-charger firmware (DIN70121/OCPP), an ESP32/FreeRTOS room display and SOLIDWORKS enclosures.

Python · TypeScript/JavaScript · C/C++ · Java · PyTorch · XGBoost · FastAPI · Next.js · MuJoCo · Linux/systemd · nginx · Tailscale

• Simulated fruit fly: a Drosophila built from the fly's mapped connectome inside MuJoCo physics, with every parameter cited and labelled measured, inferred or assumed. The goal is a fly that learns. (Research project, source private.)

Special thanks to Navid M. for inspiring me and leading me down this path. Thank you, sir.