Jose Javier Gonzalez

@jjgort

Research Scientist @DbrxMosaicAI • PhD @MIT_CSAIL • Working on LLM training and inference efficiency

San Francisco, CA josejg.com Joined January 2017

164 Following 586 Followers 69 Posts

Latest posts 5

Pinned post

Jose Javier Gonzalez @jjgort Mar 3

FlashOptim is our latest research on low precision optimizers. We came up with the right numeric tricks to halve model training memory without any quality degradation, and we are releasing the library and kernels that make all of this magic possible!

Davis Blalock Verified account @davisblalock Mar 3

🚀 Today we’re releasing FlashOptim: better implementations of Adam, SGD, etc, that compute the same updates but save tons of memory. You can use it right now via `pip install flashoptim`. 🚀

arxiv.org/abs/2602.23349

A bunch of cool ideas make this possible: [1/n]

View quoted post on X

Jose Javier Gonzalez @jjgort Aug 27

RIP this meme, it was fun while it lasted

Max ⛅ Verified account @maxisawesome538 Feb 7, 2024

when u ask HuggingFace what the business plan is

View quoted post on X

Jose Javier Gonzalez @jjgort Jul 21

I've been doing this for months, and I can confirm it's great. I go one step further and hook the cleanup/clarification into the harness (github.com/JJGO/pi-rephra…), so that it can use the full session context to resolve ambiguities and provide better test instructions.

Andrej Karpathy Verified account @karpathy Jul 21

One pattern I find useful for working with LLMs is a nice long ramble session. Sometimes the LLM needs more bits to understand what you're trying to achieve, but you're too lazy to type them. In these cases I like to lean back, switch to /voice and just ramble for like 10 minutes, total mess, anything goes, full stream of consciousness. Sometimes I declare it up top, something like "switching to speech recognition sorry for any typos...". Sometimes I turn it into a small interview of a few turns. But I find that the LLMs are somehow very good at reconstructing long incoherent rambles and often their echo of your own tangle of thoughts comes out quite a bit cleaner than what you started with. The result is that you improve the mind meld and have to correct things less from that point on.
View quoted post on X

Jose Javier Gonzalez @jjgort Jul 10

This completely matches my first hand experience, Pi is faster, more accurate, and cheaper than codex or claude code for the same task. The customization is the cherry on top. Thanks @badlogicgames for making a harness that is 3.14x better than the other ones

Matei Zaharia Verified account @matei_zaharia Jul 8

We benchmarked coding agents on our own internal tasks at Databricks and learned a lot!

There are many surprising opportunities to lower cost and increase quality, and many models including open source ones are truly competitive now. 🧵

View quoted post on X

Jose Javier Gonzalez @jjgort Jul 9

I'm at #ICML2026 this week, and today I'll be presenting FlashOptim, our latest work on memory efficient optimizers! We are in Poster Session 8 at 5pm, so come to HALL A #410 to get a free explainer, grill me with questions, or to hear the war stories of doing optimizer numerics!