Software engineer shipping AI into full-stack systems, end-to-end — from system design to deploy. Python, TypeScript, LiveKit, React.
I'm a computer engineer based in Istanbul, shipping production systems where AI is a feature — not the product. My work sits at the intersection of reliable backends, polished React frontends, and AI capabilities (voice agents, vision, embeddings, search) that actually make it to paying users.
Day-to-day I write Python (FastAPI, async SQLAlchemy, LiveKit, ElevenLabs) and TypeScript (React 19, shadcn/ui, Tailwind v4, Framer Motion), and I own features end-to-end — system design, backend, UI, CI/CD, deploy. Correctness at the token level matters to me: clean types, logical CSS, accessible primitives, predictable state.
I graduated from Marmara University and work comfortably in English, Turkish, and Arabic. I'm open to roles where I can keep owning features from design to deploy — especially in teams that value thoughtful engineering over shipped-and-forgotten.
A mix of production systems and targeted internal tools — each one owned end-to-end, from voice pipeline or backend service to the final UI.
Restaurant voice ordering platform
Multi-restaurant voice ordering platform. Customers place orders by talking to a LiveKit-powered voice agent on the restaurant's website; owners manage menus, sessions, and orders from a dashboard. Real-time agents handle turn-taking, ordering, and handoff to the staff.
Multi-tenant AI voice agent for medical tourism
AI voice platform for medical tourism clinics. Handles inbound calls, WhatsApp threads with GPT vision-based dental photo analysis, Bitrix CRM webhooks, and outbound call scheduling. Currently migrating the original single-tenant prototype into a multi-tenant platform with JWT auth, company-scoped data isolation, and per-tenant encrypted secrets.
Lab report generator with PDF parsing and CLIP image clustering
Desktop-oriented tool for a Turkish photometric testing lab. Parses multiple PDF report types with PyMuPDF, crops vector-graphic diagrams by text-anchor detection, clusters uploaded product photos via CLIP embeddings, and fills a Word template by cross-run XML replacement. Trilingual UI (EN/TR/AR with RTL).
Neighborhood marketplace
Web platform for local neighborhood markets (semt pazarı) in Istanbul — listings, vendor accounts, search by product and location. Split frontend + backend services.
A focused toolkit, picked for production reliability rather than novelty. Most of the day is TypeScript and Python; the rest is whatever the problem calls for.
2025 — present
OkurTech · Istanbul
Own features end-to-end across a team's production voice-AI platforms (Yaver, Quicall) and domain-specific tools (Report Automator): system design, Python/FastAPI services and LiveKit/ElevenLabs agents, React + shadcn frontends, plus CI/CD pipelines and deployment on Docker, Railway, and self-managed Linux.
2024 — 2025
Infosec · Saudi Arabia (remote)
Built AI-centered products and experimentation systems: voice biometric verification, call summarization, a Gemini-based translator, and teacher-support tooling. Hands-on across Python/FastAPI, React, Vue, and model pipelines (Whisper, PyAnnote, ChromaDB).
Oct 2022 — Jan 2023
Hinda Smart City Technologies · Istanbul
Transformed 120 GB of raw multi-format data (Excel, Access, SQL, .txt, .dbf, .gdb) into a centralized Elasticsearch index using Logstash pipelines. Built automated extraction/cleaning/transformation jobs and recovered previously-inaccessible legacy datasets for analysis.
Aug 2022
Sudatel Telecommunication · Khartoum, Sudan
Worked inside an agile IoT software team on telecom user-data tooling — subscriber info, SIM management, service subscriptions. Contributed to system-design discussions for scalable, secure IoT systems.
Sep 2019 — Jul 2023
Marmara University · Istanbul
Relevant coursework: Software Engineering, Data Structures, OOP, Databases, Computer Networks, Open-Source Software.
The fastest way to reach me is email. I reply within a day, usually sooner.
