Miles聽Lubin
Principal Software Engineer, Decision Intelligence at NVIDIA
New York, New York, United States
Summary
馃ぉ
Rockstar馃帗
Top SchoolMiles Lubin is a Principal Software Engineer with 14 years of experience applying mathematical optimization and automatic differentiation to production systems, currently leading decision-intelligence work on NVIDIA's cuOpt. He holds a Ph.D. in Operations Research from MIT and has bridged research and engineering roles at Google Research and Hudson River Trading, shipping algorithms that combine rigorous theory with practical performance. An active open-source contributor, Miles has made substantive backend contributions to the Julia language and key optimization libraries鈥擮ptim.jl, ForwardDiff.jl, JuMP.jl鈥攁nd helped extend solver support in CVXPY, reflecting deep expertise in convex and numerical optimization. Colleagues rely on him for performance-focused fixes (e.g., AD, Hessian sparsity, and type-stable code) and for turning complex mathematical ideas into robust, tested software.
14 years of coding experience
8 years of employment as a software developer
Doctor of Philosophy (Ph.D.), Operations Research, Doctor of Philosophy (Ph.D.), Operations Research at Massachusetts Institute of TechnologyUniversidad Nacional de C贸rdobaStaples High SchoolMS, Statistics, MS, Statistics at University of Chicago
Spanish, Portuguese, French, Hebrew, English
Github Skills (23)
autodiff10
algorithms10
python10
solver10
convex-optimization10
data-structure10
mathematical-optimization10
calculus10
automatic-differentiation10
data-structures10
julia10
optimization10
test-automation10
testing9
benchmark9
Programming languages (17)
Github contributions (5)
jump-dev/JuMP.jl
Oct 2012 - Jul 2022
Modeling language for Mathematical Optimization (linear, mixed-integer, conic, semidefinite, nonlinear)
Role in this project:
Back-end Developer
Back-end DeveloperContributions:14 releases, 233 reviews, 1739 commits in 9 years 11 months
Contributions summary:Miles primarily contributed to the optimization and improvement of the sparsity detection mechanism within the Julia language modeling framework JuMP. The commits show the user implementing a graph coloring approach to exploit Hessian sparsity and fixing existing sparsity detection methods. Furthermore, they modified and extended core functionalities such as the user-defined function and constraint processing, and addressed issues related to linear and quadratic expressions.
conic-programsmathematicallinear-programmingnonlinear-programminglinear
jump-dev/Convex.jl
Nov 2014 - Jun 2020
A Julia package for disciplined convex programming
Role in this project:Back-end Developer
Back-end DeveloperContributions:2 reviews, 33 commits, 28 PRs in 5 years 7 months
Contributions summary:Miles primarily contributed to the core functionalities of the Julia package `convex.jl`, which involves disciplined convex programming. Their commits focused on updating solver capabilities, modifying tests, and fixing compatibility issues related to Julia versions. Specifically, the changes included standardizing types, updating solver information, and adapting the code to work with newer versions and formats of the project's dependencies.
julia-packageconvex-optimizationsolverconvexjulia
Find and Hire Top DevelopersWe鈥檝e 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.