HuggingFace
Developer Platforms, LLM APIs, and AI Assistants with a focus on practical implementation and code.
Nutrition Label
This channel provides high-fidelity technical content for developers, ranging from live coding sessions to architectural breakdowns of modern AI models. The videos excel at authenticity, frequently showing real-time terminal workflows, debugging steps, and reproducible code rather than abstract slides. While deep dives offer rigorous mathematical context, shorter product updates tend to showcase streamlined 'happy path' implementations.
Strengths
- +
- +
- +
Notes
- !Deep dives explore mathematical theory and edge cases, while product updates focus on basic functionality.
- !Tutorials rely on first-party libraries; check documentation for integration with external tools.
Rating Breakdown
Breakdown across the key dimensions we rate. Methodology →
Recent Videos
Build a PyTorch ReLU Kernel with Hugging Face Kernels (CPU + Metal)
How to use Claude Code to automate model training IN MINUTES
From AI Agents to Faster Kernels: Ben Burtenshaw & Felix LeClair (AI Plumbers #2)
Talk: Kernels Deep Dive (Ben Burtenshaw)
Benchmarking LLMs at the Game Of Science (Eleusis)
Hugging Face Journal Club: GLM-5: from Vibe Coding to Agentic Engineering
How to create your own custom conversation app on Reachy Mini 🤖💬
Kimi K2.5 vs Claude Code (REAL Use Cases): New KING of Coding??
MoE Token Routing Explained: How Mixture of Experts Works (with Code)
Training Dashboards with Trackio + Hugging Face
OpenAI Agents SDK Crash Course (with Hugging Face Models)
Reachy Mini at Nvidia's Jensen CES keynote
Building Agents with Smolagents
How to contribute to Open Source - 7 EASY steps 🤗
HuggingChat | Chat with Open Models
Why this rating
Evidence receipts showing why each dimension is rated the way it is.
“So today I'm going to show you how to create your own conversation application.”[00:55] →
The video delivers exactly what is promised in the title: a step-by-step guide to building a custom app for the specific hardware shown.
“So I'm going to create a file called repro.py... and I'm going to try to reproduce the bug.”[02:15] →
The presenter demonstrates the critical engineering practice of creating a reproduction script to confirm a bug before attempting to fix it, showing the code running live.
“Dropping Oversubscribed Tokens”[16:51] →
The video explicitly addresses a critical edge case/limitation in MoE models (capacity constraints) rather than just explaining the happy path.