Rohan Yadav
Computer Science PhD Student at Stanford University
Pittsburgh, Pennsylvania, United States
Summary
🤩
Rockstar🎓
Top SchoolRohan Yadav is a Computer Science PhD student at Stanford University with a decade of experience building high-performance backend systems, compilers, and parallel runtimes. He has production experience on Cockroach Labs' SQL Execution team and has contributed to notable open-source projects like the Tensor Algebra Compiler (taco) and the Legion parallel programming system, where he improved robustness, tests, and low-level runtime behavior. His work spans industry (Facebook, Uber ATG) and academia (CMU), combining compiler/runtime optimizations—down to implementing OpenMP runtime calls and fixing subtle fusion/variable bugs—with a focus on maintainability. He also mentors students as a long-time teaching assistant, bringing practical engineering rigor to both research and teaching.
10 years of coding experience
3 years of employment as a software developer
Bachelor’s Degree, Computer Science, Minor in Machine Learning, Bachelor’s Degree, Computer Science, Minor in Machine Learning at Carnegie Mellon UniversityHigh School, High School at North Allegheny Senior High School
English, Hindi
Github Skills (21)
c-language10
parallel10
parallelization10
algebra10
sparse-matrix10
mappings10
parallel-processing10
openmp10
parallel-computing10
tensor10
sparse-data10
c-programming-language10
runtimes9
run-time9
sys9
Programming languages (14)
Github contributions (5)
tensor-compiler/taco
Jan 2021 - May 2022
The Tensor Algebra Compiler (taco) computes sparse tensor expressions on CPUs and GPUs
Role in this project:
Back-end Developer
Back-end DeveloperContributions:18 reviews, 303 commits, 51 PRs in 1 year 4 months
Contributions summary:Rohan primarily contributed to fixing compilation warnings and improving the code's robustness by addressing issues related to variable references within the tensor algebra compiler. They also deduplicated dimension-checking routines, improving code maintainability and reducing redundancy. Additionally, they addressed a bug in the fuse transformation, resolving undefined variable issues during compilation. Finally, they added and improved test cases.
cudatacolinear-algebratensor-algebra-compilergpu
StanfordLegion/legion
Aug 2021 - Dec 2022
The Legion Parallel Programming System
Role in this project:Back-end Developer
Back-end DeveloperContributions:13 commits, 262 comments, 83 issues in 1 year 4 months
Contributions summary:Rohan primarily contributed to the Legion Parallel Programming System, focusing on improvements within the default mapper, example programs, and the realm/openmp modules. They fixed error messages in the default mapper and added an example for backpressuring task execution using a custom mapper. Furthermore, the user implemented `omp_get_num_places` and `omp_get_num_procs` runtime calls in Realm's OpenMP API, and fixed local task registrations.
system-programminglegionparallel-programmingparallel
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.