Michael Chen

Hi! I'm a senior at CMU, on leave this semester working with Prof. Max Simchowitz on VLMs for spatial reasoning. Previously, I was lucky to intern at Cartesia where I studied H-Nets and next-latent prediction. I've also worked with Prof. Graham Neubig on LLM performance prediction and interned at AWS. I like machine learning and baseball.

Some things I've worked on:

• Gated Linear Attention in Needle for 10-714. I wrote GLA in my own deep learning library with both the recurrent and chunkwise algorithms. My chunkwise implementation reached 100x speedup over recurrent.
• Not-Just-Scaling Laws: Towards a Better Understanding of the Downstream Impact of Language Model Design Decisions (EMNLP 2025)