Lyndon Duong
Machine Learning Vision Scientist at Apple
San Francisco Bay Area United States
Summary
👤
Senior🎓
Top SchoolLyndon Duong is a Machine Learning Vision Scientist in the Bay Area with 11 years of experience bridging computational neuroscience, software engineering, and video compression research. He holds a PhD from NYU where his thesis on adaptive coding and stochastic representational geometry was advised by Eero Simoncelli and David Heeger, and he has translated that theory into practice at Google (PhD intern) by building adaptive lossy autoencoders that yielded a ~25% BD-Rate improvement over AV1. Now at Apple, he works at the intersection of human and computer visual perception and ML for video, combining deep research instincts with production-grade C++/TensorFlow implementation skills. An active contributor to numerics-heavy open-source (e.g., implementing robust L1/L2 norms and optimized Eigen::Map templates in the Stan Math library), he brings uncommon fluency in both mathematical algorithm design and low-level library engineering.
11 years of coding experience
2 years of employment as a software developer
Master of Science - MS, Physiology and Pharmacology, Master of Science - MS, Physiology and Pharmacology at Western UniversityDoctor of Philosophy - PhD, Neuroscience (Computational & Theoretical), Doctor of Philosophy - PhD, Neuroscience (Computational & Theoretical) at New York UniversityBachelor of Science - BS, Physiology and Physics, Bachelor of Science - BS, Physiology and Physics at McGill University
Github Skills (15)
mathematics10
mathlib10
automatic-differentiation10
c-language10
stan10
eigen10
cprogramming-language10
math10
math-library10
cpp-templates9
linear-algebra9
html-template9
python-templates9
apptemplates9
code-templates9
Programming languages (5)
Github contributions (5)
stan-dev/math
Dec 2021 - Jan 2022
The Stan Math Library is a C++ template library for automatic differentiation of any order using forward, reverse, and mixed modes. It includes a range of built-in functions for probabilistic modeling, linear algebra, and equation solving.
Role in this project:
Back-end Developer / Library Developer
Back-end Developer / Library DeveloperContributions:5 reviews, 25 commits, 4 PRs in 13 days
Contributions summary:Lyndon's primary contributions involve implementing and testing mathematical functions within the Stan Math Library, specifically focusing on L1 and L2 norm calculations. They've developed both prim and rev versions of the `norm1` and `norm2` functions, creating unit tests to ensure their correctness and handle potential edge cases, such as NaN values. Furthermore, they adapted the `dot_self` function and made templating and Eigen::Map implementations to optimize for various input types, demonstrating a strong understanding of numerical algorithms and C++ template programming.
automatic-differentiationprobabilistic-modelingsundialsmodestemplate-library
lyndond/lyndond.github.io
Dec 2017 - Dec 2022
Personal site
Contributions:111 commits, 2 PRs, 119 pushes in 5 years
tricksmozillareact
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.