Working with such a top-notch team of researchers and engineers at @lagrangedev is inspirational. Excited to see DeepProve being used in the real world and cannot wait to see what's next! Kudos to our CEO @Ismael_H_R for being a constant source of encouragement from day 1!
AI doesn’t need to be a black box.
So we opened ours.
DeepProve is now open source. The fastest zkML in the world.
Fork it. Run it. Break it. Build something weird.
→ https://t.co/TaVk7RUNHr
6/ None of this would have happened without the researchers and engineers at @lagrangedev who poured themselves into it. Thank you @nikkolasg1, @Ismael_H_R, Tianyi Liu, Nicholas Mainardi, @dipapadopo, Christodoulos Pappas, Shravan Srinivasan, Zack Youell, and @YupengZhang7!
1/ After 1.5 years of focused work with a dedicated team at @lagrangedev, I am thrilled to announce that our central project, DeepProve, has been accepted at CCS 2026 and is now fully open-source. See paper here https://t.co/Fygm1BfnmO and code here https://t.co/rf0mOuJO47.
5/ Performance: DeepProve proves inference at ~174 tokens/min (GPT-2) and ~86 tokens/min (Gemma 3), while verification takes just 1 to 3.7 seconds. A distributed implementation scales throughput substantially further.
The team at Lagrange has gone and done it again!
We're very excited that "Efficiently Provable Approximations for Non-Polynomial Functions" has been accepted to the USENIX Security Symposium.
I'm brewing a new course component on Crypto <> AI for my fall course: https://t.co/bNFdmkd2Dx, featuring several works from IC3.
I would love to incorporate a few talks/interviews with people building in this area. If you are interested, please DM me!
Intereresting take on “consensuslessness” for payments by @poddotnetwork researcher @giorgos_tsimos. When applications do not require agreement on everything, you can agree faster.
Building decentralized markets requires tackling challenges related to efficiency, real-time latency, market fairness, and censorship-resistance.
Running black-box consensus at each step of the way won’t cut it.
@abresas and @giorgos_tsimos discuss how Pod is utilizing a dual-architecture and enforceable market rules to provide fast and fair markets on decentralized networks on a global scale.
🎙️New episode of Inside the Pod 👇
We’re excited to see three papers from our research interns accepted to @ZKProof VIII in Rome.
◆ Dynamic zk-SNARKs
◆ Relaxed Modular PCS from Arbitrary PCS and Applications to SNARKs for Integers
◆ Code-based Scalable Collaborative SNARKs
Work like this pushes forward the foundations of verifiable computation
Congrats to Weijie, Alireza, and Chris.
Alireza
Alireza, another research intern, had his paper on relaxed modular PCS accepted to @ZKProof in Rome:
"What inspired this work was the gap between proof systems built over finite fields and the fact that real-world statements are often over integers. We tried to bridge that gap."
One more step toward making proving practical for real-world applications.
Congrats to my 7th PhD student, Weijie Wang, on successfully defending "Dynamic Zero-Knowledge Proofs & their Applications"! His work on updatable SNARKs, sparse arguments, and recursion-free IVC will be presented at EUROCRYPT 2026: https://t.co/B1Mxnj2OUI. Bravo Dr. Wang!
🔐 Announcing @SeQureDB2026 — the 1st Workshop on Secure & Private Data Management, co-located with @SIGMODConf 2026 in Bengaluru!
Organized by @SujayaMaiyya, @JennieRogers & me.
Let's bridge the gap between crypto/security ↔ databases/systems and academia ↔ industry. 👇 (1/3)