Alexandre Eichenberger
Research Scientist at IBM T.J Watson Research Lab.
Greater New York City Area United States
Summary
🤩
Rockstar🎓
Top SchoolAlexandre Eichenberger is a research scientist with 11 years of experience specializing in high-performance compilers and parallel computing for multi-core, SIMD and GPU architectures. He combines deep academic training from ETH Zurich and the University of Michigan with practical systems work at IBM, focusing on compilation for AI workloads such as ONNX and scalable execution via OpenMP and software pipelining. His contributions to the ONNX-MLIR backend include shape inference, code generation for element-wise operations and robust memory management, plus unit tests that improved correctness and performance. Known for low-level loop transformations, scheduling and auto-parallelization, he translates architecture-aware theory into production-ready compiler optimizations. Based in the Greater New York City area, he brings a rare mix of formal research rigor and hands-on engineering that accelerates ML model deployment on heterogeneous hardware.
11 years of coding experience
Ph.D, Computer Science & Engineering, Ph.D, Computer Science & Engineering at University of MichiganDiploma, Computer Science, Diploma, Computer Science at ETH Zurich
German, French
Github Skills (15)
tensorrt10
mathematics10
unit-testing10
memory-management10
compiler-design10
code-generation10
c-language10
inference10
tensor10
tensorflow10
cprogramming-language10
math10
shapes10
operation10
maths10
Programming languages (9)
Github contributions (5)
onnx/onnx-mlir
Mar 2020 - Jan 2023
Representation and Reference Lowering of ONNX Models in MLIR Compiler Infrastructure
Role in this project:
Back-end Developer
Back-end DeveloperContributions:1 release, 1953 reviews, 239 commits in 2 years 11 months
Contributions summary:Alexandre primarily contributed to the implementation of core functionalities within the ONNX-MLIR compiler, focusing on the integration of mathematical operations, notably implementing shape inference and code generation for concatenation and other element-wise unary operations. They also worked on memory management and low-level optimization, as demonstrated by their work on memory allocation and memory copy operations. Furthermore, they improved the accuracy of the code and fixed some related bugs by implementing a robust unit testing for all the implemented operation.
pytorchrepresentationdeep-learningmlironnx-models
AlexandreEichenberger/onnx-mlir
Mar 2020 - Mar 2025
Contributions:2111 pushes, 348 branches in 5 years 1 month
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.