Lagrange’s internship program offers rolling admission for exceptional researchers and engineers eager to work at the frontier of cryptography. Interns contribute directly to the development of leading zero-knowledge proof systems and verifiable compute mechanisms applied across AI, defense, and blockchain. Participants gain access to mentorship from top researchers and the full suite of Lagrange's resources to transform ideas into real-world impact.
Yale University
Researched a fair zero-knowledge marketplace for provers and clients, formulating a double-auction mechanism that achieves truthfulness, weak group-strategy-proofness, and near-optimal social welfare.
2024 Internship Cohort
PhD Candidate, Hong Kong University of Science and Technology
Collaborated on the design and development of DeepProve, Lagrange’s verifiable AI system, and conducted research toward the first scalable, transparent, and post-quantum collaborative zkSNARK.
2025 Internship Cohort
University of California, Berkeley
Developed a general technique for efficiently provable approximations of non-linear functions, enabling accurate zk-proofs for exponential, trigonometric, and other continuous operations used in zkML.
2025 Internship Cohort
University of Pennsylvania
Designed a lattice-based modular polynomial commitment scheme for SNARKs on integer arithmetic, advancing post-quantum-secure proof systems that natively support non-field computations.
2025 Internship Cohort
.avif)
Charalampos (Babis) Papamanthou is the Chief Scientist at Lagrange and the director of Lagrange's internship program. A professor of Computer Science at Yale and Co-Director of the Yale Applied Cryptography Laboratory, his research spans verifiable computation, privacy-preserving systems, and blockchain security—earning support from the NSF, JP Morgan, and Protocol Labs. With over 15,000 citations and pioneering contributions that have shaped modern cryptography, Papamanthou brings a visionary, hands-on approach to building the infrastructure of verifiable AI and secure computation.
DeepProve proved inference for Google’s Gemma3, unveiling a unified Einsum layer, optimized tensor commitments, and a distributed graph architecture that accelerates zkML at scale.
October 20, 2025
