Education
From podcasts to midnight tinkering, we’re all about getting hands-on. It’s about sparking curiosity and turning big ideas into actual code.
Twinkle Podcast
Real-world developer stories and lessons learned from the trenches.
- EP00 | Learning LoRA the Hard Way - Ah-Qian
- EP01 | The Real Life of a Freelance Developer: Challenges & Achievements - Phidias
- EP02 | From Knowledge Fragments to Deep Exchange: The Origin of AILogora - ChiChieh
- EP03 | The Rise of Voice AI: Making Machines Understand and Respond to You - Ethan
Midnight Tinkering
10 PM. Fans spinning. Might as well start training.
A collaborative workshop at its core, this space is where theory meets raw implementation through building datasets and training small models from the ground up.
Dismantling the process to see what’s under the hood is the heartbeat of every session — because when the fans are spinning, sleep is a luxury for people without GPUs.
- GitHub Repository: ai-twinkle/llm-lab
(Dedicated labs for Traditional Chinese model research)
Midnight Deep Dives
Stripping It Down to the Studs
In this reading group, moving beyond just flipping pages means taking theory apart to see what makes it tick. Systematically deconstructing "Build a Large Language Model (From Scratch)" provides a real handle on the foundations before ever writing a single line of code.
Translating dense technical concepts into digestible, community-driven notes happens in every meetup. It’s about clearing the path from raw theory to actual understanding, ensuring the underlying logic is transparent for everyone involved.