Artem L.

I build automation with n8n, LLMs and local AI models: business workflows with LLM agents, media pipelines on ComfyUI, and batch automation for 3D and Houdini. Before automation: 12+ years in 3D and nearly a decade in Houdini, the last five building production pipelines for synthetic data. From that work I keep the same approach: runs without supervision, error handling and logs, the same output from the same input. What I build: - n8n workflows with LLM agents: processing requests, documents and messages, routing, enrichment, integrations between services, with error handling and monitoring - AI media pipelines: ComfyUI image and video generation behind n8n or a bot, from prompt preparation to the finished file - Telegram and Slack bots as the front end to automation - Local setups with Ollama and ComfyUI on your own hardware, when data must not leave it - Multi-agent runs with Claude Code for code and data audits - Automation for 3D: PDG batch generation, headless Houdini runs, Houdini connected to ComfyUI and LLMs Examples: - A Telegram bot that edits photos and turns images into video: a Flask server, n8n orchestration, ComfyUI and Ollama running locally, Slack notifications, a separate workflow that catches errors. - An automated video production pipeline: configuration in Google Docs, scripts written by an LLM, visuals from ComfyUI, local text-to-speech, assembly, all orchestrated in n8n. Demos and work samples available on request.