Research
My work focuses on code generation and scaling deep learning systems. At MSL, I help lead code-generation research across FAIR and TBD, supporting a team of scientists and engineers in Menlo Park, New York, and Paris.
Some active and past areas of research include:
- Code generation: large-scale code-generation research; previously co-led Code World Model (CWM)
- Multimodal generative models: co-led Chameleon and built Transfusion
- Speech recognition: led Libri-Light and built wav2letter
I've also led the development of open-source machine learning software, including Flashlight, Shumai, and wav2letter, parts of which were integrated into PyTorch.
Selected Invited Talks
- CWM: Building World Models for Computation - AI Engineer (2025) (talk)
- Scaling Asynchronous RL Recipes for Learning to Code - SC 2025, Meta (2025) (slides)
- Trends in Deep Learning Computation and Frameworks - Intel (2024)
- Multimodal Generative Models - Charles River Ventures & Greylock (2024)
- GPU Computing at Scale - University of Pennsylvania, CIS 565: GPU Programming (2024)
- Shumai: Fast, Flexible Machine Learning in TypeScript - Google DeepMind (2023)
- Flashlight: Enabling Innovation in Tools for Machine Learning - ICML (2022) (talk, slides)
- Scaling Deep Learning for Automatic Speech Recognition - NVIDIA GPU Technology Conference (2020)
Background
I studied computer science at Penn Engineering and statistics/operations research at The Wharton School through the M&T Program. I also completed graduate studies in computer and information science at Penn.
While at Penn, I ran PennApps, the largest collegiate hackathon in the world at the time.
I grew up outside of Chicago.
Teaching
I created and teach CIS 5690: GPU Programming for Machine Learning Systems at Penn.
I was a teaching assistant for several computer science courses while studying at Penn, including:
- CIS 380, 548, 700: Operating Systems (Spring/Fall 2017, 2018) - Head TA - Concurrency, resource management, virtual memory, file systems, virtual machines.
- CIS 505: Distributed Systems (Spring 2018) - Synchronization, communication, replication. Project-based, in C++.
- CIS 320: Algorithms (Spring/Fall 2017, 2018) - Head TA - Combinatorial optimization, sorting, hashing, graphs, complexity theory.
- CIS 240, 593: Introduction to Computer Architecture (Fall 2015, 2016) - Hardware structures, organization, machine language, C.
- CIS 550: Database and Information Systems (Spring 2017) - Structured data, modeling, architecture, and distributed processing.