Arya Grayeli (@AryaGrayeli) / X

Arya Grayeli

74 posts

Arya Grayeli

@AryaGrayeli

chief scientist @getsapien

Joined June 2018

Pinned
Arya Grayeli
@AryaGrayeli
Sep 17, 2024
Why doesn't search leverage common sense? Why not use learned patterns to make search more efficient? Making every step marginally more informed compounds into massive boosts. My new paper with @atharva_sehgal @swarat @MilesCranmer applies this for scientific discovery
23K
Arya Grayeli
@AryaGrayeli
Oct 29, 2024
Excited to show the world what we have been up to! Building this product and team @sapienai has pushed what we thought to be possible for AI in enterprise finance. I can't even imagine what the next few years look like as our team further grows. Stay tuned!
Ron Nachum
@RonNachum
Oct 29, 2024
Can AI run a company? We raised an $8.7M seed led by @generalcatalyst and top tech & finance angels to prove it can. At @sapienai, we’re building autonomous financial analysts redefining enterprise finance—already saving thousands of hours and catching multi-million-$ mistakes
00:00
2.9K
Arya Grayeli
@AryaGrayeli
Sep 25, 2024
Excited to announce that we have been accepted into @NeurIPSConf 2024! Can't wait to see everyone in Vancouver! Drop any recs or DM me if you want to meet up there #neurips2024
Arya Grayeli
@AryaGrayeli
Sep 17, 2024
Why doesn't search leverage common sense? Why not use learned patterns to make search more efficient? Making every step marginally more informed compounds into massive boosts. My new paper with @atharva_sehgal @swarat @MilesCranmer applies this for scientific discovery
4.8K
Arya Grayeli
@AryaGrayeli
Dec 13, 2024
It was great to present at NeurIPS this week! My research in neurosymbolic programming introduces learning a concept library to accelerate search. We use this @sapienai to learn internal company concepts and their unique practices. If you’re around Vancouver and want to chat
7.3K
Arya Grayeli
@AryaGrayeli
Jul 14, 2024
A little preview on what’s to come! 👀
Ali Partovi
@apartovi
Jul 14, 2024
I'm excited to start the @Neo Accelerator 1-mo bootcamp in Colorado! Even before it begins, I ran into the Sapien founders @RonNachum @PranavRavella and @AryaGrayeli -- on the jetway to board the plane haha. 🤩
999
Arya Grayeli
@AryaGrayeli
Oct 29, 2024
Replying to @RonNachum @generalcatalyst and @sapienai
building with these guys is the best every day. hyped for whats to come!
705
Arya Grayeli
@AryaGrayeli
Jun 21, 2024
Looking at @Kalshi markets during a break, this contract just spiked from 43 to 60 on the @AnthropicAI drop today. Tons of volume on the contract and using Sonnet today was amazing. Hard buy still? Seems like sellers are thinking GPT-5 is dropping soon.
kalshi.com
Odds & Predictions
Track what Kalshi's markets predict for "GPT no longer top ranked LLM", or trade it yourself.
2.2K
Arya Grayeli
@AryaGrayeli
Jul 11, 2024
The community at Neo are some of the most supportive, genuine, and bright people I have met. Ali and the Neo family really value the younger generation in tech and have been a huge part of our journey. Join the fam!
Ali Partovi
@apartovi
Jul 9, 2024
Calling future tech leaders: Neo Scholar applications are open. Apply now! Past @neo Scholars are crushing it in AI and beyond. (🧵...)
1.3K
Arya Grayeli
@AryaGrayeli
Jul 26, 2024
On the surface, @GoogleDeepMind solving 4/6 IMO problems is great, but searching through solution candidates then proving them is hard to scale. No wonder it can't solve the two combinatorics problems. Everything is a search problem but is a more direct approach necessary?
666
Arya Grayeli
@AryaGrayeli
Sep 18, 2024
my guys are killing it!
Brendan (can/do)
@BrendanFoody
Sep 18, 2024
Mercor is solving global labor matching with models that understand human ability. @mercor_ai raised a $30M Series A at a $250M valuation, led by @victoralazarte and @bgurley at @benchmark, with participation from @peterthiel, @jack, @adamdangelo, and @LHSummers.
00:00
683
Arya Grayeli
@AryaGrayeli
Jul 29, 2024
Lack of proper eval methods is the biggest problem in ML today. Can't wait for a world where benchmarks are truly informative and not overfit. Love the direction @scale_AI is taking.
Alexandr Wang
@alexandr_wang
Jul 29, 2024
1/ Scale is announcing our latest SEAL Leaderboard on Adversarial Robustness! 🛡️ Red team-generated prompts 🎯 Focused on universal harm scenarios 🔍 Transparent eval methods SEAL evals are private (not overfit), expert evals that refresh periodically scale.com/leaderboard
1.6K
Arya Grayeli
@AryaGrayeli
Jul 17, 2024
There's something about coding with classical music on wired headsets. Broke my AirPods so using my old wired earbuds and the act of plugging in and putting on an old playlist is nostalgic. Been absolutely locked in
365
Arya Grayeli
@AryaGrayeli
Sep 3, 2024
Cursor is my favorite prompt engineer
659
Arya Grayeli
@AryaGrayeli
Sep 17, 2024
Replying to @AryaGrayeli
Huge thanks to the entire team throughout this process and @mlfoundry for the unlimited compute. See the paper here: arxiv.org/pdf/2409.09359
571