Lab Home | Phone | Search
Center for Nonlinear Studies  Center for Nonlinear Studies
 Home 
 People 
 CNLS Staff Members 
 Executive Committee 
 Postdocs 
 Visitors 
 Students 
 Research 
 Publications 
 Conferences 
 Workshops 
 Sponsorship 
 Talks 
 Seminars 
 Postdoc Seminars Archive 
 Quantum Lunch 
 Quantum Lunch Archive 
 P/T Colloquia 
 Archive 
 Ulam Scholar 
 Anastasio Fellow 
 
 Student Requests      
 Student Program 
 Visitor Requests 
 Description 
 Past Visitors 
 Services 
 General 
 
 History of CNLS 
 
 Maps, Directions 
 T-Division 
 LANL 
 
Tuesday, September 09, 2025
09:00 AM - 11:00 AM
T-4 Conference Room (TA03, Building 524, Room 105)

Seminar

Catalyst: just-in-time compilation of structured quantum programs

Josh Izaac
Xanadu

Up until now, most quantum programming frameworks have been written in Python, and serialize solely the quantum part of the workflow to simplistic string-based representations that are sent to cloud-connected quantum hardware (while the classical part executes locally!). But this ignores the history of classical programming infrastructure, and the fact that no algorithm is purely quantum — there is bound to be expensive and interwoven classical processing, and we need to take this into account.

In this talk, he’ll chat about how they are addressing this with PennyLane via Catalyst, a framework for quantum just-in-time compilation (QJIT). Using QJIT, full quantum-classical programs written in Python are automatically captured and compiled using standard compiler technologies such as MLIR and LLVM — leading to not only performance improvements, but increasingly richer ways we can interface with quantum computers.

Using this approach, he’ll then show how they can perform constant-time hybrid compilation of Shor’s algorithm; he will demonstrate that with QJIT compilation, the algorithm is compiled once per bit width of N, the integer being factored, even when N-specific optimizations are applied to circuit generation based on values determined at runtime. The implementation is benchmarked up to 32-bit, and both the size of the compiled program and the pure compilation time are found to be constant (under 3 seconds on a laptop computer), meaning code generation becomes tractable even for realistic problem sizes.

Bio: Josh Izaac holds a Ph.D. in graph algorithms for quantum computing, with a focus on software development and computational physics, from the University of Western Australia Following completion in 2017, he then moved to Toronto to take up a position at Xanadu, working across education, research, content, and software development. As Head of Quantum Software, he helped develop Xanadu’s open-source quantum software ecosystem, including Strawberry Fields, an open-source Python library for programming Xanadu’s early continuous-variable photonic hardware; PennyLane, an open-source differentiable quantum programming framework; and Catalyst, a just-in-time compiler for hybrid quantum-classical programs. Since 2022 he has been the Director of Product at Xanadu, leading high-level roadmap planning and strategy for Xanadu’s open-source quantum programming ecosystem across software, content, and cloud services, with the aim of enabling breakthrough research through novel software.

Host: Kasia, Krzyzanowska, MPA-Q