Prithviraj (Raj) Ammanabrolu
@rajammanabrolu
RL and Language. Assistant Prof @UCSanDiego. Research Scientist @Nvidia.
The PEARLS Lab at @UCSD_CSE is now open for business! I'm recruiting Fall 24 PhD students in all things interactive and grounded AI, RL, and NLP!! Join us in the land of 🏖️ beach (🧋pearl tea included). Apply by Dec 20. Please help spread the word! More: pearls.ucsd.edu
Soon™, I'll be an Asst Prof @UCSanDiego @UCSD_CSE focusing on interactive & grounded AI, RL, NLP I will also be a research scientist @MosaicML helping lead efforts to make tech like RLHF more accessible Looking for PhD students & research eng/scientists to join me in ☀️SoCal🏖️
Saw a bunch of self driving type things being tested (?) Amusing to see that they were all following speed limits and getting a crazy number of ppl beeping (at things that can't respond) Wonder if companies will make Socal self driving RL policies or if humans here will adapt?
(Inland?) Socal is crazy. I'm going 90-100+ and still barely keeping up with the traffic. Starting to think public transport here basically has no future
.@ccui9 will be doing a contributed talk on our recent TALES: Text Adventure Learning Environment Suite here. Lots of progress to be made in RL envs for more general reasoning! Go check it out and say hi! arxiv.org/abs/2504.14128
Our #ACL2025NLP workshop REALM on LLM agents is happening July 31 in Vienna 🎶🎼 🗓️ Schedule & accepted papers are live! realm-workshop.github.io 🚀Join us for a day of invited talks, paper presentations and a panel discussion with an amazing line-up!
For a few years now, whenever I want actual quality human data for any of my projects, my first call is to @echen @HelloSurgeAI. Never been disappointed. (And yes I've sampled other offerings on the market)
$1BN+ in revenue. $0 in funding. “I would not sell to Zuck for $100BN.” Surge AI. The best company in tech that you might not have heard of. Their Founder, Edwin, never does interviews. Today he shares all with 20VC and my top 7 takeaways 👇
Senior researchers whine about how junior ones are too negative when reviewing and how peer review is broken but don't take 5 sec to check their public goal post moving reactions anytime something new happens. Where do you think the junior ones get it from? Lack of self awareness
Recent work has seemed somewhat magical: how can RL with *random* rewards make LLMs reason? We pull back the curtain on these claims and find out this unexpected behavior hinges on the inclusion of certain *heuristics* in the RL algorithm. Our blog post: tinyurl.com/heuristics-con…
Name a more iconic duo: North American conferences and visa issues
Sadly, both @hsanchaita and I will be missing @icmlconf (due to visa reasons) for our Oral presentation, but catch @TuhinChakr presenting our position paper. If you have questions, thoughts, or follow-ups, please don't hesitate to send them our way! 📧 paper & review:…
RLEF
I replicated this result, that Grok focuses nearly entirely on finding out what Elon thinks in order to align with that, on a fresh Grok 4 chat with no custom instructions. grok.com/share/c2hhcmQt…
Problems like PDF extraction and computer use agents to automate legacy software should and will die. They are bandaid solutions that slow down progress towards better designed (human centered) AI native software generally
I often rant about how 99% of attention is about to be LLM attention instead of human attention. What does a research paper look like for an LLM instead of a human? It’s definitely not a pdf. There is huge space for an extremely valuable “research app” that figures this out.