hi! i'm nathan

my life goal is to explore the potentialities of intelligence

right now i'm doing ml research at kimi moonshot, where i work on llm pretraining (i'm proud of kimi k3). my research mostly focuses on model architectures (past example) and optimizers. i also like thinking about efficient, hardware-aligned algorithms and kernels (co-design them!).

prior to this, i worked on applied interp research at tilde research. outside of work, i try to learn as much as i can.

here are some things i believe in:

• agency is a superpower
• taste is learnable
• write things down
• everything is figureoutable with friends
• the world is hackable in wholesome ways
• you'd probably love ssh config with named hosts

here are some things i like:

• serendipitous blogs, open-source, mangoes, cold showers, excalidraw, doc.new, elegant kernels, asking llms first, mixing lego pieces, constrained geometric shapes, meditating, learning from my sensei gemini (this didn't age well), intuition, light mode, cli, blocks, "twitter" - the name, twitter's rec algo

things i don't like:

• twitter's rec algo

i love people and write back. send me a message! (email probably works best)