Leland Mcinnes
Research Mathematician And Data Scientist
Ottawa, Ontario, Canada
Summary
🤩
Rockstar🎓
Top SchoolLeland Mcinnes is a research mathematician, data scientist, and software engineer with 11+ years building high-performance ML and HPC systems from laptops to supercomputers. He bridges deep theoretical work in algebraic geometry, topology and category theory with practical, optimized implementations—most notably as the lead author of the widely used UMAP dimensionality-reduction library and the pynndescent nearest-neighbor engine. At the Tutte Institute he leads a small research group focused on unsupervised learning, clustering, outlier detection and scalable graph construction, while contributing performance-critical Cython and numba optimizations across open-source projects. His background includes cryptologic research and industrial data-visualization work, which together sharpen his focus on bringing abstract mathematics to concrete, production-ready tools. Leland’s code emphasizes sparse-data handling and practical robustness, and he often prototypes rapidly in scripting environments before scaling to HPC languages. Based in Ottawa, he combines rigorous academic training (PhD in Mathematics) with a track record of production-grade ML libraries used by the community.
11 years of coding experience
2 years of employment as a software developer
Doctor of Philosophy (Ph.D.), Mathematics, Doctor of Philosophy (Ph.D.), Mathematics at Western UniversityBachelor of Science (B.Sc.) with First Class Honours, Mathematics, Bachelor of Science (B.Sc.) with First Class Honours, Mathematics at University of Canterbury
English, Japanese
Stackoverflow
346reputation
40kreached
11answers
0questions
Github Skills (42)
algorithm10
data-visualizations10
code-optimization10
algorithms10
scipy10
python10
imap10
clustering-algorithm10
scikit10
sparse-matrix10
machine-learning10
hdbscan10
data-visualisation10
data-structure10
numpy10
Programming languages (10)
Github contributions (5)
lmcinnes/umap
Jul 2017 - Jan 2023
Uniform Manifold Approximation and Projection
Role in this project:
Back-end Developer
Back-end DeveloperContributions:31 releases, 15 reviews, 1134 commits in 5 years 7 months
Contributions summary:Leland contributed to the UMAP library's core functionality by implementing and optimizing key algorithms for dimension reduction. They focused on the performance of several aspects of the project by implementing specialized routines in Cython, and in particular, contributed significantly to the sparse matrix functionality. Furthermore, this user appears to be in charge of creating several base level algorithms for other applications.
projectiondimensionality-reductionmachine-learningtopological-data-analysisapproximation
scikit-learn-contrib/hdbscan
Apr 2015 - Nov 2022
A high performance implementation of HDBSCAN clustering.
Role in this project:Data Scientist
Data ScientistContributions:52 releases, 11 reviews, 756 commits in 7 years 7 months
Contributions summary:Leland's commits center around the addition and optimization of tensor mutual reachability distance calculations within a machine learning context. They modified the code to better represent the algorithm and its implementation. Furthermore, they implemented cluster scoring and the addition of soft clustering, which demonstrates an understanding of clustering evaluation and analysis.
clusteringhdbscanmachine-learningmachine-learning-algorithmsclustering-algorithm
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.