👋 I'm Cristian Dorobănțescu — AI automation engineer and full-stack developer based in Romania.
I build custom AI workflows (LLMs, image/video generation, speech synthesis) and medical web applications (membership platforms, conference systems, patient registries) for organizations that need reliable systems. I published the first F5-TTS model fine-tuned for Romanian and run my own Proxmox infrastructure.
Most ComfyUI tutorials show you how to generate a single image. I'll show you how to wire it into n8n, handle errors gracefully, and produce 200 consistent images for a client without touching the UI.
Feb 27, 2026Medical clients don't care about clever architecture. They care about the system being up on congress day, the export working before the deadline, and not having to call you at 11pm.
Feb 27, 2026Cloud GPUs are expensive if you leave them running. Here's the hybrid setup I use: Proxmox for everything persistent, RunPod for burst inference, and Tailscale to make it feel like one machine.
Feb 27, 2026Romanian has almost no open-source TTS options. I spent three months fixing that — collecting data, hitting GPU walls, and eventually publishing the first F5-TTS model fine-tuned on Romanian speech.
Feb 27, 2026First F5-TTS speech synthesis model fine-tuned for Romanian. Fills a significant gap in open-source Romanian TTS — trained on native speech data, published on Hugging Face.