Josh Pollock
PHD Student at Paul G. Allen School of Computer Science & Engineering
Cambridge, Massachusetts, United States
Summary
🤩
Rockstar🎓
Top SchoolJosh Pollock is a PhD student and software researcher with nine years of experience building systems at the intersection of programming languages, formal methods, and deep learning compilers. Based at MIT after undergraduate work at the University of Washington, he has contributed to high-profile open-source projects such as Apache TVM—improving Python package integration and adding Relay operators—and led full-stack improvements to the Penrose diagramming engine. His background includes research on program state visualization and transpilation (Coq2Lean), plus internships in formal verification and hardware-software integration at Apple and Intel. Josh combines rigorous academic training with practical engineering chops, often tackling build/integration and constraint-layout problems that sit between theory and production.
9 years of coding experience
Mathematics and Computer Science, Mathematics and Computer Science at University of WashingtonHigh School, High School at Lick-Wilmerding High SchoolHigh School, High School at Aragon High SchoolHigh School, High School at Stanford Online High SchoolDoctor of Philosophy - PhD, Doctor of Philosophy - PhD at Massachusetts Institute of Technology
Github Skills (22)
javascript10
visualization10
python10
mathematics10
typescript10
diagram10
math10
relayer10
typescript-types10
typescripts10
visualizations10
relay10
compiler-compiler9
functional-programming9
compiler9
Programming languages (14)
Github contributions (5)
penrose/penrose
Mar 2020 - Jul 2021
Create beautiful diagrams just by typing notation in plain text.
Role in this project:
Full-stack Developer
Full-stack DeveloperContributions:33 reviews, 8 commits, 7 PRs in 1 year 4 months
Contributions summary:Josh made significant contributions to the Penrose project, primarily focused on improving the diagram generation process. They fixed issues related to continuous map examples and incorporated arrowheads into path lengths within the rendering engine. Furthermore, the user refactored the constraint system, unifying bounding box computations and improving the overall layout of the project. They also implemented style-controlled canvas dimensions, ensuring proper rendering of diagrams.
diagramssvgeulermathematicsjavascript
apache/tvm
Jul 2018 - Jun 2019
Open deep learning compiler stack for cpu, gpu and specialized accelerators
Role in this project:Back-end Developer
Back-end DeveloperContributions:18 commits, 22 PRs, 276 comments in 10 months
Contributions summary:Josh primarily contributed to the build and integration of Python packages and dependencies within the TVM repository. They made updates to the installation scripts, specifically focusing on adding and managing required Python packages like `orderedset` and `antlr4`, which indicates an involvement in setting up the environment for development and testing. They also implemented new Relay operators and modified existing ones, adding features like a clip operator, which demonstrates a focus on the core functionality of the deep learning compiler.
metalvulkancompilertensoropencl
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.