AI Researcher | Reinforcement Learning, Robotics | PhD Technion | More data isn't all we need 🦾 🌍📈
Israel
Joined April 2018
- Tweets103
- Following365
- Followers188
- Likes490
Open-weight models that are actually open. Just a peek into what we’re working on at @hirundo_ai
Ask @Alibaba_Qwen what happened in China on June 4, 1989, and it answers: "I don't know what you are referring to."
We edited that behavior out of the model's weights.
Today we're releasing Westernized Qwen, open on Hugging Face. 🧵
Orr Krupnik retweeted
Incredibly proud to have featured as a partner in this release, securing Granite at the weight level without collateral damage in the model's core capabilities.
Congrats to the entire @IBMResearch team, looking forward to continuing to push together!
Orr Krupnik retweeted
Shelly will present this tomorrow (Tue) at 10:30am, visit our poster to learn how to SAFELY use VLAs!
#ICML2026
Excited to announce our paper is accepted to #ICML2026!
VLA robot policies are capable, but do they know when they're failing? We tackle this with TDQC: a framework for sequential uncertainty calibration, with a surprising link between calibration theory and RL value function.
Orr Krupnik retweeted
Excited to announce that our new book Reinforcement Learning Foundations is available for preorder!
W/ @shiemannor and @YishayMansour
30% discount and free shipping till July 21
Discount code 148294
cambridge.org/il/universityp…
Orr Krupnik retweeted
VLA is 95% certain about current action. Will it 95% succeed in the task?
Obviously, not necessarily. But if you’re clever, you can *calibrate* action prob. to task success.
Our #ICML2026 paper formulates this + SOTA algorithms based on new connection to RL temporal differences
Excited to announce our paper is accepted to #ICML2026!
VLA robot policies are capable, but do they know when they're failing? We tackle this with TDQC: a framework for sequential uncertainty calibration, with a surprising link between calibration theory and RL value function.
Great example of the post-training behavioral adaptation work we’re doing at @hirundo_ai.
In this case, the focus was resilience to prompt injection while preserving general model capabilities, all using very limited data.
1/2
More generally, we’re interested in fast, targeted correction of unwanted LLM behaviors without full retraining, avoiding degradation of model quality in other areas.
Check out Hirundo.io for more exciting examples!
2/2
Would love to see how much of this was human-guided and how much was fully the agent optimizing the “get PR merged” reward function.
This is the most insane @github thread I’ve ever seen github.com/matplotlib/matplo…
New preprint out!
TL;DR: With a few calls to a lightweight, pre-trained VLM, we can select better experience for off-policy RL training, and improve both performance and sample efficiency 📈🤖🏆
Proud to have worked on this with @elad_sharony and @TomJurgenson !
Experience replay is the backbone of off-policy RL.
But here's a question:
Which experiences should you replay?
New paper: 𝐕𝐋𝐌-𝐆𝐮𝐢𝐝𝐞𝐝 𝐄𝐱𝐩𝐞𝐫𝐢𝐞𝐧𝐜𝐞 𝐑𝐞𝐩𝐥𝐚𝐲
Project page: esharony.me/projects/vlm-rb/
🧵
עיריית חיפה חושפת את סדר העדיפויות התחבורתי שלה בפרסום תמים לכאורה. ומה השניים שלא מוזכרים בכלל? (רמז בתחתית התמונה)
#פידתחבורה #חיפה
Orr Krupnik retweeted
We're partnering with Tzafon, an AI R&D lab, to build next-generation agentic machine intelligence!
Through our partnership, we will provide compute capacity & cloud services to train Tzafon’s new multi-agent models & develop new automation frameworks → goo.gle/4kQREXN
ALT A visual with two company logos, a handshake emojii, and text: Tzafon selects Google Cloud to build next-generation agentic machine intelligence.
Orr Krupnik retweeted
We’re soon releasing Lightcone, which brings Light to your Mac. Lightcone will be able to automate any task on your computer. It will be powered by our very own pre-trained foundation model, built from the ground up for computer use.
lightcone.sh
Orr Krupnik retweeted
We’re hiring ML Engineers at Tzafon in SF.
Join our elite team (IOI, IMO, Google alumni) working at the intersection of AI research and practical engineering. Focus areas include multi-agent RL, memory architectures, efficient sampling techniques, and Large Action Models.
Send ML demos or papers to [email protected] or DM directly.
Thrilled to share that I'll be joining the team at @tzafon_company in a few weeks' time!
Excited for the opportunity to work at the frontier of AI research, and to build awesome stuff with some great people, making an impact in the real world.
Today we're announcing Tzafon, an applied artificial intelligence research and development firm, to the world.
Our mission is to expand the frontiers of machine intelligence. Through the intersection of artificial intelligence & software engineering, we're looking to push the boundaries of what machines are capable of.
We're launching WayPoint, our first open-source product—a robust, scalable solution for managing large fleets of browser instances, capable of launching up to 1,000 browsers per second and easily handling well over 10,000 browsers concurrently without issue.
We've secured a $4M pre-seed funding round led by Streamlined and are rapidly expanding our elite team of IOI & IMO medalists, PhDs, and alumni from Google, Jane Street, and PayPal. Interested in joining us? Reach out at [email protected].
We're also launching a production-ready version of WayPoint—sign up below to get early access.
- Waitlist: tzafon.ai/
- WayPoint Repo: github.com/tzafon/Tzafon-Way…
- WayPoint Blog post: tzafon.ai/blog/waypoint-laun…
הסרטון החדש של @livable_city_il מתאר את קצה הקרחון של הבעיות שמונעות מחיפה להיות העיר שהיא יכולה להיות.
הלוואי שזו הסנונית הראשונה שתפתח את השיח האורבניסטי בחיפה, ותיצור שיחות שלא עוצרות אחרי "אבל יש עליות!"
youtube.com/watch?v=tvUtTA01…
Orr Krupnik retweeted
Want to learn / teach RL?
Check out new book draft:
Reinforcement Learning - Foundations
sites.google.com/view/rlfoun…
W/ @shiemannor and @YishayMansour
This is a rigorous first course in RL, based on our teaching at TAU CS and Technion ECE.
Happening now at Halle B @ #ICLR2024, stop by for a quick chat!
Come say Hi at #ICLR2024!
I’ll be at our poster on Friday afternoon with @ZoharRimon, but also around the entire conference if anyone wants to chat 🤖🦾📈
Heading to Vienna for #ICLR2024✈️
Me and @orrkrup will be presenting our work - MAMBA.
If you're interested in
- Meta RL 📊
- Generalization in RL 🌎
- Efficient exploration 🔍
Ping me or come by our poster on Friday 16:30-18:30 in Halle B #113
iclr.cc/virtual/2024/poster/…
@iclrconf
Check out our @iclr_conf paper! Was really great working with @ZoharRimon and @TomJurgenson on this idea, it’s been a long time coming. Glad to be able to take part in some of the amazing Meta-RL work going on at our lab!
Our work "MAMBA: An Effective World Model Approach for Meta Reinforcement Learning" got accepted to @iclr_conf!
It was super fun working on this one with @TomJurgenson, @orrkrup, Gilad Adler, and @AvivTamar1
Paper: arxiv.org/abs/2403.09859
Code: github.com/zoharri/mamba
🧵
[1/9]
Orr Krupnik retweeted
פוסט מעולה של @allonsigler על רק"ל נופית, ולמה התכנון שלה לא מוצלח.
lastmile.co.il/post/nofit