Ahmad Salim Al-Sibahi
Research Scientist at University of Copenhagen
Copenhagen, Capital Region of Denmark
Summary
🤩
RockstarAhmad Salim Al-Sibahi is a research scientist and assistant professor based in Copenhagen with 11 years of experience at the intersection of machine learning and probabilistic modeling. He combines academic rigor with hands-on engineering—contributing to high-profile open-source ML tooling such as Pyro—where he implemented von Mises sampling via the Best-Fisher algorithm and improved inference and predictive modules. His work shows deep expertise in probability distributions, sampling techniques, and debugging complex inference code paths. Comfortable moving between research and production, he focuses on making probabilistic methods more reliable and user-friendly for practitioners. Colleagues value his blend of meticulous algorithmic thinking and practical software craftsmanship.
11 years of coding experience
Github Skills (17)
pytorch10
variational-inference10
probabilistic-programming10
python10
data-modeling10
machine-learning10
probabilistic-reasoning10
bayesian10
statistical-models10
probabilistic-models10
bayesian-inference10
deep-learning9
data-structures8
data-structure8
algorithms8
Programming languages (20)
Github contributions (5)
pyro-ppl/pyro
Nov 2018 - Jan 2020
Deep universal probabilistic programming with Python and PyTorch
Role in this project:
ML Engineer
ML EngineerContributions:8 commits, 9 PRs, 75 comments in 1 year 3 months
Contributions summary:Ahmad contributed to the core functionality of the probabilistic programming library, Pyro. They implemented sampling for the von Mises distribution using the Best-Fisher algorithm, showcasing expertise in probability distributions and sampling techniques. The user also fixed bugs and improved the error messages related to log probabilities and score computations within the inference and predictive modules, demonstrating a strong understanding of the library's internal workings. Furthermore, they improved the TracePredictive module and added new features to keep specific sites.
pytorchpythondeep-learningbayesian-inferenceprobabilistic-modeling
ahmadsalim/numpyro
Apr 2020 - Nov 2020
Probabilistic programming with NumPy powered by JAX for autograd and JIT compilation to GPU/TPU/CPU.
Contributions:139 commits, 112 pushes, 9 branches in 7 months
cpucompilationpythonjitautomatic-differentiation
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.