I build AI systems people actually talk to — LLM assistants, real-time voice agents, and the full-stack products around them.
I'm a software and AI engineer who works across the whole stack — from training and orchestrating models to shipping the interfaces people use them through. My focus is LLM systems: assistants, voice agents, and retrieval pipelines that hold up in production.
I care about latency, reliability, and products that feel simple even when the system underneath isn't.
Conversational assistants, real-time voice agents, RAG pipelines, evals, and guardrails — designed for production, not demos.
Complete web products — from data model and API design to polished, fast frontends. I ship the whole thing.
Model training and deployment, streaming pipelines, and the infra that keeps AI products fast and observable.
Real-time guest-service agent for hotels — handles requests, questions, and concierge tasks the moment a guest asks.
Conversational assistant for hotels — guest requests, bookings, and concierge questions, grounded on each property's data.

E-commerce website for a jewellery brand — product catalog, collections, and online storefront.
NFC-tagged digital passports for animals — vaccination records, medical history, and identity, readable with a tap. In production with 500+ veterinarians across Turkey.