blog

notes on efficient inference, multi-token generation, and multimodal machine learning

Thought on TTT-E2E

Just read the TTT-E2E paper and had some informal thoughts.

1 min read · January 13, 2026

2026 · thought
Recent Trend in Architectures

A recurring recipe seems to be “more parallel computation paths, each narrower.”

2 min read · January 06, 2026

2026 · thought
Published 4 Papers!

A quick update on four 2025 papers.

1 min read · December 04, 2025

2025 · update
November & December 2025 Reading List

This November and December, I was interested on MoE efficiency, speculative decoding, and efficient attention methods.

6 min read · December 04, 2025

2025 · blog
Discrete Diffusion for Text Infilling

Project page and paper for flexible-length text infilling with discrete diffusion models.

1 min read · October 21, 2025

2025 · thought
Multimodal Paper Reviews

Paper reviews from my graduate multimodal class.

52 min read · March 25, 2025

2025 · blog