My name is Luan, and I'm Vietnamese.
My favorite topics are efficient diffusion models and language models.
I've spent several years working as an AI engineer. My most recent position was Head of Decentralized AI at ZenAI-Vietnam, where I led a small team. At ZenAI, I work on decentralized ai model serving, as a subnet of Bittensor.
After resigned from ZenAI, I once tried launching a startup focused on decentralized LLM accerleration 🚀, in a collaboration with Taostats Validator. Though it didn't ultimately succeed but I learnt a lot.
I'm passionate about Machine Learning, spanning a wide range of topics from high-level algorithms to precision engineering. I use PyTorch for GPU interactions, Ray for distributed serving, and I can write simple yet efficient Triton kernels.
Currently, I'm in a career transition, dedicating my time to deepening my knowledge of theory from classic to modern generative models, relearning the math I didn't fully explore during my engineering career, immersing myself in research papers for greater insight into the field, and contributing to open-source projects.
Open Source Contributions: