Devan Shah

Hi! I am Devan Shah, a recent graduate of Princeton, where I am very grateful to have been advised by Prof. Elad Hazan. I am broadly interested in information flow within LLMs, accelerating inference of State Space Models, and LLMs for mathematics.

Some notable things worth mentioning:

• Thesis in convex training & fast inference of SSMs (SpectraLDS, FutureFill).
• Excited by the problems and promise of LLM super-human teaching.
• Recipient of the Phillip Goldman '86 Senior Prize (Princeton's top CS honor) and the Lore von Jaskowsky Memorial Prize (Princeton SEAS's top research honor).

I’ve had the pleasure of interning and researching at Google DeepMind, Jane Street, and TikTok/ByteDance. Outside research, I’ve been honored to lead Princeton ACM.

Feel free to reach out at devan.shah [at] princeton [dot] edu.