Sergio Charles

I’m a co-founder of Darwin, where we build general-purpose robots for assembling and maintaining the physical infrastructure of civilization, starting with AI infrastructure. Previously, I worked on AI R&D at Google X, NVIDIA, and the Stanford AI Lab, including autonomous RL for robotics with Chelsea Finn and computer vision with Andrew Ng. I earned a B.S. in Mathematics and Computer Science and an M.S. in Statistics from Stanford. My background spans pure math, reinforcement learning, and natural language processing.

latest posts

Mar 21, 2025	Quantum Mechanics
Dec 28, 2024	Protein Function Prediction using LoRA-Relational Graph Neural Networks and Multi-Task Learning
Aug 7, 2023	A Neat Derivative Trick

selected research

2025

Black Box to Bedside: Distilling Reinforcement Learning for Sepsis Time Series

2025

Bib HTML

Leveraging deep learning models to increase the representation of nomadic pastoralists in health campaigns and demographic surveillance

2025

Bib HTML

@article{sergio_nomadic_pastoralist,
  title = {Leveraging deep learning models to increase the representation of nomadic pastoralists in health campaigns and demographic surveillance},
  author = {},
  journal = {},
  year = {2025},
  publisher = {Stanford},
}

2024

FuncE GNN: Protein Function Prediction using Multi-Task and Relational Graph Learning

Sergio Charles

2024

Bib HTML

@article{sergio_funcegnn,
  title = {FuncE GNN: Protein Function Prediction using Multi-Task and Relational Graph Learning},
  author = {Charles, Sergio},
  journal = {},
  year = {2024},
  publisher = {Stanford},
}

Spacetime E(n)-Transformer: Equivariant Attention for Spatio-temporal Graphs

Sergio Charles

2024

Bib HTML

@article{sergio_spacetime,
  title = {Spacetime E(n)-Transformer: Equivariant Attention for Spatio-temporal Graphs},
  author = {Charles, Sergio},
  journal = {},
  year = {2024},
  publisher = {Stanford},
}

2021

Loss Agnostic and Model Agnostic Meta Neural Architecture Search for Few-Shot Learning

Sergio Charles, Gordon Chi, and Gil Kornberg

CS229, 2021

Bib HTML

@article{sergio_metalearning,
  title = {Loss Agnostic and Model Agnostic Meta Neural Architecture Search for Few-Shot Learning},
  author = {Charles, Sergio and Chi, Gordon and Kornberg, Gil},
  journal = {CS229},
  year = {2021},
  publisher = {Stanford},
}

Variable Quantization Noise for Neural Network Compression

Sergio Charles, William Ellsworth, and Lyron Co Ting Keh

CS224N, 2021

Bib HTML

@article{sergio_nlp,
  title = {Variable Quantization Noise for Neural Network Compression},
  author = {Charles, Sergio and Ellsworth, William and Keh, Lyron Co Ting},
  journal = {CS224N},
  year = {2021},
  publisher = {Stanford},
}

2019

The Existence of Infinitely Many Geometrically Distinct Non-Constant Prime Closed Geodesics on Riemannian Manifolds

Sergio Charles

math.DG, 2019

Bib HTML

@article{sergio_geodesics,
  title = {The Existence of Infinitely Many Geometrically Distinct Non-Constant Prime Closed Geodesics on Riemannian Manifolds},
  author = {Charles, Sergio},
  journal = {math.DG},
  year = {2019},
  publisher = {Arxiv},
}

2018

On the Time Evolution of Ricci Scalar Curvature in the Late Epoch for a Λ-CDM-Parameterized Universe

Sergio Charles

phys, 2018

Bib HTML

@article{sergio_gr,
  title = {On the Time Evolution of Ricci Scalar Curvature in the Late Epoch for a {$\Lambda$}-CDM-Parameterized Universe},
  author = {Charles, Sergio},
  journal = {phys},
  year = {2018},
}

2017

On the Algebro-Geometric Analysis of Meromorphic (1,0)-forms

Sergio Charles

math.DG, 2017

Bib HTML