Cole Medin
youtubeAI engineering YouTuber focused on agents, RAG, and local LLM workflows.
Cole Medin began publishing AI-agent content on YouTube in 2023, initially focused on the early wave of GPT-4-based assistants and quickly pivoting to the agent-framework space as LangChain, LlamaIndex, and later Pydantic AI matured. His background is in software engineering; the channel reflects an engineer-first sensibility with deep walk-throughs of code, environment setup, and the failure modes people actually hit when productionising AI agents.
Medin's content style is unusually structured for the AI-creator space. Tutorials are typically 30 to 60 minutes long, scripted with code samples on-screen, and accompanied by a public GitHub repository linked in the description. He has covered topics including Pydantic AI agent construction, multi-agent orchestration with LangGraph, vector-database selection for RAG, integrating local LLMs via Ollama, and using n8n as an AI-agent runtime. Several of his repos accumulate substantial star counts independent of the videos.
His audience skews technical: software engineers, data engineers, and AI/ML practitioners who want hands-on detail rather than executive overviews. Channel growth accelerated through 2024-2025 as the AI-agent topic matured into a recognised engineering discipline, reaching approximately 100,000 subscribers by December 2025. Medin's present focus, based on his 2025-2026 publishing cadence, is the intersection of local LLM deployment, MCP (Model Context Protocol) integration, and self-hosted agent infrastructure.
Our Take
We point clients' engineering teams toward Medin's tutorials when they need a working reference architecture rather than a vendor demo. The Pydantic AI and LangGraph series in particular have saved us multiple weeks of internal experimentation on agent projects. — Rafal Fila, ShadowGen
What Sets Cole Medin Apart
Medin is one of the few AI-content creators who consistently ships runnable open-source code alongside each video. The depth of integration coverage (Pydantic AI, LangGraph, n8n, vector stores, local LLMs) and the engineer-first format distinguish him from the more business-focused side of AI YouTube.
Key Achievements
Topics Covered
Content Style
Long-form tutorials (30-60 minutes) with on-screen code, accompanied by public GitHub repositories. Engineer-first explanation depth rather than executive overview.
Standout Content
His 'n8n AI Agent from Scratch' tutorial is the definitive guide to building AI agents with n8n. Over 35 minutes, he builds a complete AI agent that can search the web, access a knowledge base, and take actions — all within n8n. He covers the vector database setup, prompt engineering, tool configuration, and error handling that other tutorials skip. If you want to build production AI agents with n8n, this is where you start.
Notable Content
Content Schedule
Weekly to twice-weekly long-form tutorials
Who Watches Cole Medin
Software engineers, data engineers, and AI practitioners building production agent systems. Audience skews hands-on technical, not buyer-side.
Work With Cole Medin
Collaborates with developer-focused tools for technical reviews and integration tutorials. Open to open-source project partnerships and developer community events.