Shoaib Kamil
Principal Research Scientist at Adobe
Cambridge, Massachusetts, United States
Summary
🤩
Rockstar🎓
Top SchoolShoaib Kamil is a Principal Research Scientist based in Cambridge, MA with 15 years of experience building high-performance compilers and graphics runtimes. At Adobe he leads research that bridges systems-level performance engineering and compiler IR design, drawing on prior research roles at MIT and Lawrence Berkeley National Lab. He contributes to prominent open-source projects—most notably Halide and the Tensor Algebra Compiler (taco)—where his work on OpenGL runtime fixes and IR node design improved portability and code-generation for data-parallel and sparse-tensor workloads. Shoaib combines deep PhD-level systems knowledge with practical production optimizations, often focusing on subtle runtime state and codegen issues that yield outsized performance gains.
15 years of coding experience
11 years of employment as a software developer
Doctor of Philosophy (PhD), Computer Science, Doctor of Philosophy (PhD), Computer Science at University of California, Berkeley
English, Spanish
Stackoverflow
1reputation
0reached
1answer
0questions
Github Skills (22)
c-language10
algebra10
glsl10
intermediate-code10
compiler-design10
performance-optimization10
irr10
tensor10
cprogramming-language10
opengl10
code-generator9
code-generation9
compiler-compiler9
compiler9
gpu8
Programming languages (6)
Github contributions (5)
tensor-compiler/taco
Oct 2016 - Nov 2019
The Tensor Algebra Compiler (taco) computes sparse tensor expressions on CPUs and GPUs
Role in this project:
Back-end Developer
Back-end DeveloperContributions:67 commits, 5 PRs, 16 pushes in 3 years 1 month
Contributions summary:Shoaib primarily focused on developing the internal representation (IR) and related components of the Tensor Algebra Compiler. Their work involved defining IR nodes for various operations like literals, variables, addition, and subtraction, as well as implementing a visitor pattern. They added support for additional IR nodes (multiplication, division, etc.), an IR printer, and a function node, expanding the compiler's capabilities. The user also made improvements to the component types used by the compiler.
cudatacolinear-algebratensor-algebra-compilergpu
halide/Halide
May 2016 - Jul 2022
a language for fast, portable data-parallel computation
Role in this project:Back-end Developer & Performance Engineer
Back-end Developer & Performance EngineerContributions:73 reviews, 72 commits, 96 PRs in 6 years 3 months
Contributions summary:Shoaib primarily contributed to the `halide/halide` repository by addressing bugs and optimizing the OpenGL runtime. They fixed a bug in the texture wrapping function and improved the pipeline execution by saving and restoring OpenGL state. Furthermore, the user implemented fixes to the code generation of integer immediates in GLSL and refactored to correctly load/store texture data within the runtime.
computationhexagonhalideparallelgpu
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.