Kyle Daruwalla
NeuroAI Scholar at Cold Spring Harbor Laboratory
Madison, Wisconsin, United States
Summary
🤩
Rockstar🎓
Top SchoolKyle Daruwalla is a NeuroAI scholar and recent PhD graduate from UW–Madison with 11 years of experience at the intersection of computer architecture, machine learning, computational neuroscience, and programming languages/compilers. Now at Cold Spring Harbor Laboratory, he focuses on computational neuroscience and unconventional, low-power online learning paradigms inspired by biological mechanisms like Hebbian learning and STDP. His research and engineering background spans academic research, hardware-focused internships at Texas Instruments and AMD, and practical ML tooling—he contributed precise output-dimension calculations and tests to the popular Flux.jl deep learning library. Kyle combines hands-on digital design and verification experience with a deep theoretical grounding, enabling him to translate neuro-inspired algorithms into efficient architectures and compiler-aware implementations. Colleagues describe him as someone who bridges messy hardware constraints and elegant learning models, often surfacing simple, testable fixes that improve usability and reliability.
11 years of coding experience
6 years of employment as a software developer
Doctor of Philosophy (PhD), Electrical Engineering, Doctor of Philosophy (PhD), Electrical Engineering at University of Wisconsin-MadisonBachelor’s Degree, Computer Engineering / Continuous Applied Mathematics, 3.65/4.00, Bachelor’s Degree, Computer Engineering / Continuous Applied Mathematics, 3.65/4.00 at Rose-Hulman Institute of Technology
Stackoverflow
925reputation
17kreached
18answers
5questions
Github Skills (19)
artificial-neural-networks10
machine-learning10
deeplearning-ai10
deep-learning10
flux10
neural-network10
fluxor10
julia10
data-structures8
data-structure8
algorithms8
algorithm8
data-science7
scala-compiler6
integer6
Programming languages (13)
Github contributions (5)
FluxML/Flux.jl
Dec 2019 - Sep 2022
Relax! Flux is the ML library that doesn't make you tensor
Role in this project:
ML Engineer
ML EngineerContributions:626 reviews, 191 commits, 71 PRs in 2 years 9 months
Contributions summary:Kyle contributed to the Flux.jl library by adding functionality for calculating the output dimensions of various layers. They implemented the `outdims` methods for several convolutional, pooling, and basic layers, including `Conv`, `ConvTranspose`, `DepthwiseConv`, `CrossCor`, `MaxPool`, `MeanPool`, `Dense`, and `Maxout`. These changes involved calculating the output sizes based on input dimensions, strides, padding, and kernel sizes, which is important for building and understanding neural network architectures. The user also added tests to verify the accuracy of these output dimension calculations, ensuring the library's usability for deep learning model design.
ml-librarythe-human-braindata-sciencedeep-learningneural-networks
FluxML/ParameterSchedulers.jl
Dec 2020 - Oct 2022
Common hyperparameter scheduling for ML
Contributions:12 reviews, 164 commits, 67 PRs in 1 year 10 months
schedulinghyperparameteroptimizationhyperparametersmachine-learning
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.