Alp Dener
Senior Performance Engineer at NVIDIA
Chicago, Illinois, United States
Summary
🤩
Rockstar🎓
Top SchoolAlp Dener is a Senior Performance Engineer with 13 years of experience specializing in numerical optimization, high-performance computing, and ML framework performance tuning. With a PhD in aerospace engineering and a research pedigree at Argonne and RPI, he bridges rigorous applied-math methods and production-grade distributed/heterogeneous code—authoring PETSc optimiser enhancements and GPU-accelerated quasi-Newton prototypes. At Graphcore and NVIDIA he shipped and benchmarked MLPerf workloads, optimized Transformer training, and contributed to Transformer Engine and DL framework performance pipelines. He combines deep expertise in adjoint- and matrix-free methods with hands-on CUDA and low-precision training work, making him adept at squeezing performance from both scientific codes and modern deep learning stacks.
13 years of coding experience
5 years of employment as a software developer
Doctor of Philosophy (Ph.D.) Aerospace Aeronautical and Astronautical Engineering, Doctor of Philosophy (Ph.D.) Aerospace Aeronautical and Astronautical Engineering at Rensselaer Polytechnic InstituteUMBC
English, Turkish
Stackoverflow
65reputation
6kreached
0answers
2questions
Github Skills (6)
c1710
numerical-optimization10
c1110
linear-algebra10
performance-optimization9
boost-python6
Programming languages (7)
Github contributions (5)
petsc/petsc
Feb 2018 - Dec 2022
Mirror of https://gitlab.com/petsc/petsc
Role in this project:
Back-end Developer
Back-end DeveloperContributions:247 commits in 4 years 10 months
Contributions summary:Alp appears to have made code changes in the TaoSolver and LMVM matrix implementations, suggesting they were involved in optimizing the performance or adding new features to the numerical optimization algorithms implemented in the PETSc/PETSc repository. They also changed interfaces for least squares implementations within the TAO (Toolkit for Advanced Optimization) to include new API calls for residual and Jacobian computation, alongside the addition of the Gauss-Newton algorithm to the least squares problem class. These changes were targeted around adding functionality to the codebase, likely for increased algorithm performance.
petscgitlab
denera/TransformerEngine
Sep 2023 - Jan 2025
A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit floating point (FP8) precision on Hopper and Ada GPUs, to provide better performance with lower memory utilization in both training and inference.
Contributions:3 PRs, 463 pushes, 51 branches in 1 year 4 months
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.