Hi! I am Alper Erten, a machine learning engineer, with expertise across deep learning, natural language processing, and computer vision.

Hi! I am Alper Erten, a machine learning engineer, with expertise across deep learning, natural language processing, and computer vision.

In my prior life, I led a team of engineers helping companies reduce their carbon footprint through energy modeling.

In my prior life, I led a team of engineers helping companies reduce their carbon footprint through energy modeling.

Linkedin

Github

Resume

Deep Learning

Deep Learning

Semi-supervised image classification

This study explores the application of the FixMatch algorithm to lightweight image classifiers, aiming to optimize semi-supervised learning for resource-constrained environments. Leveraging the CIFAR-10 dataset, the study evaluates the performance of three architectures: ResNet-8, ResNet-20, and a custom OptVGGNet model. The findings highlight the potential of lightweight models in achieving efficient learning while reducing computational demands and data-labeling costs in semi-supervised setups.

PYTORCH

Semi-supervised image classification

This study explores the application of the FixMatch algorithm to lightweight image classifiers, aiming to optimize semi-supervised learning for resource-constrained environments. Leveraging the CIFAR-10 dataset, the study evaluates the performance of three architectures: ResNet-8, ResNet-20, and a custom OptVGGNet model. The findings highlight the potential of lightweight models in achieving efficient learning while reducing computational demands and data-labeling costs in semi-supervised setups.

PYTORCH

Semi-supervised image classification

This study explores the application of the FixMatch algorithm to lightweight image classifiers, aiming to optimize semi-supervised learning for resource-constrained environments. Leveraging the CIFAR-10 dataset, the study evaluates the performance of three architectures: ResNet-8, ResNet-20, and a custom OptVGGNet model. The findings highlight the potential of lightweight models in achieving efficient learning while reducing computational demands and data-labeling costs in semi-supervised setups.

PYTORCH

Image Style Transfer

This paper identifies two important axioms that effective attribution methods should satisfy which are “sensitivity” and “implementation invariance”. Without an axiomatic approach, it is difficult to determine whether an attribution method is affected by artifacts from the data, network architecture, or the attribution method itself. Existing attribution methods fail to simultaneously satisfy both axioms. As a solution, the paper introduces the Integrated Gradients method, which satisfies both.

PYTORCH

Image Style Transfer

This paper identifies two important axioms that effective attribution methods should satisfy which are “sensitivity” and “implementation invariance”. Without an axiomatic approach, it is difficult to determine whether an attribution method is affected by artifacts from the data, network architecture, or the attribution method itself. Existing attribution methods fail to simultaneously satisfy both axioms. As a solution, the paper introduces the Integrated Gradients method, which satisfies both.

PYTORCH

Image Style Transfer

This paper identifies two important axioms that effective attribution methods should satisfy which are “sensitivity” and “implementation invariance”. Without an axiomatic approach, it is difficult to determine whether an attribution method is affected by artifacts from the data, network architecture, or the attribution method itself. Existing attribution methods fail to simultaneously satisfy both axioms. As a solution, the paper introduces the Integrated Gradients method, which satisfies both.

PYTORCH

Machine Learning

Machine Learning

Unsupervised Learning and Dimensionality Reduction

In this article, we will study the application of two clus- tering algorithms and three linear dimensionality reduction algorithms on two selected classification datasets through a series of experiments.

PYTORCH

Unsupervised Learning and Dimensionality Reduction

In this article, we will study the application of two clus- tering algorithms and three linear dimensionality reduction algorithms on two selected classification datasets through a series of experiments.

PYTORCH

Unsupervised Learning and Dimensionality Reduction

In this article, we will study the application of two clus- tering algorithms and three linear dimensionality reduction algorithms on two selected classification datasets through a series of experiments.

PYTORCH

Supervised Learning

This report uses two datasets to explore the behavior and performance of three machine learning algorithms: Support Vector Machines (SVM), k-Nearest Neighbors (KNN), and Neural Networks (NN) with two selected dataset. The ex- ploration mainly focuses on the model training and tuning process as well as the comparison of final model prediction performance and the computational costs associated with each algorithm.

PYTORCH

Supervised Learning

This report uses two datasets to explore the behavior and performance of three machine learning algorithms: Support Vector Machines (SVM), k-Nearest Neighbors (KNN), and Neural Networks (NN) with two selected dataset. The ex- ploration mainly focuses on the model training and tuning process as well as the comparison of final model prediction performance and the computational costs associated with each algorithm.

PYTORCH

Supervised Learning

This report uses two datasets to explore the behavior and performance of three machine learning algorithms: Support Vector Machines (SVM), k-Nearest Neighbors (KNN), and Neural Networks (NN) with two selected dataset. The ex- ploration mainly focuses on the model training and tuning process as well as the comparison of final model prediction performance and the computational costs associated with each algorithm.

PYTORCH

Robotics

Robotics

Manufacturing robotics

This study explores the application of the FixMatch algorithm to lightweight image classifiers, aiming to optimize semi-supervised learning for resource-constrained environments. Leveraging the CIFAR-10 dataset, the study evaluates the performance of three architectures: ResNet-8, ResNet-20, and a custom OptVGGNet model. The findings highlight the potential of lightweight models in achieving efficient learning while reducing computational demands and data-labeling costs in semi-supervised setups.

PYTORCH

Manufacturing robotics

This study explores the application of the FixMatch algorithm to lightweight image classifiers, aiming to optimize semi-supervised learning for resource-constrained environments. Leveraging the CIFAR-10 dataset, the study evaluates the performance of three architectures: ResNet-8, ResNet-20, and a custom OptVGGNet model. The findings highlight the potential of lightweight models in achieving efficient learning while reducing computational demands and data-labeling costs in semi-supervised setups.

PYTORCH

Manufacturing robotics

This study explores the application of the FixMatch algorithm to lightweight image classifiers, aiming to optimize semi-supervised learning for resource-constrained environments. Leveraging the CIFAR-10 dataset, the study evaluates the performance of three architectures: ResNet-8, ResNet-20, and a custom OptVGGNet model. The findings highlight the potential of lightweight models in achieving efficient learning while reducing computational demands and data-labeling costs in semi-supervised setups.

PYTORCH

App Development

App Development

Android development

I made an android app

PYTORCH

Android development

I made an android app

PYTORCH

Android development

I made an android app

PYTORCH

Energy storage & carbon emissions tool

This study explores the application of the FixMatch algorithm to lightweight image classifiers, aiming to optimize semi-supervised learning for resource-constrained environments. Leveraging the CIFAR-10 dataset, the study evaluates the performance of three architectures: ResNet-8, ResNet-20, and a custom OptVGGNet model. The findings highlight the potential of lightweight models in achieving efficient learning while reducing computational demands and data-labeling costs in semi-supervised setups.

PYTORCH

Energy storage & carbon emissions tool

This study explores the application of the FixMatch algorithm to lightweight image classifiers, aiming to optimize semi-supervised learning for resource-constrained environments. Leveraging the CIFAR-10 dataset, the study evaluates the performance of three architectures: ResNet-8, ResNet-20, and a custom OptVGGNet model. The findings highlight the potential of lightweight models in achieving efficient learning while reducing computational demands and data-labeling costs in semi-supervised setups.

PYTORCH

Energy storage & carbon emissions tool

This study explores the application of the FixMatch algorithm to lightweight image classifiers, aiming to optimize semi-supervised learning for resource-constrained environments. Leveraging the CIFAR-10 dataset, the study evaluates the performance of three architectures: ResNet-8, ResNet-20, and a custom OptVGGNet model. The findings highlight the potential of lightweight models in achieving efficient learning while reducing computational demands and data-labeling costs in semi-supervised setups.

PYTORCH

Portfolio level carbon emissions reduction tool

This study explores the application of the FixMatch algorithm to lightweight image classifiers, aiming to optimize semi-supervised learning for resource-constrained environments. Leveraging the CIFAR-10 dataset, the study evaluates the performance of three architectures: ResNet-8, ResNet-20, and a custom OptVGGNet model. The findings highlight the potential of lightweight models in achieving efficient learning while reducing computational demands and data-labeling costs in semi-supervised setups.

PYTORCH

Portfolio level carbon emissions reduction tool

This study explores the application of the FixMatch algorithm to lightweight image classifiers, aiming to optimize semi-supervised learning for resource-constrained environments. Leveraging the CIFAR-10 dataset, the study evaluates the performance of three architectures: ResNet-8, ResNet-20, and a custom OptVGGNet model. The findings highlight the potential of lightweight models in achieving efficient learning while reducing computational demands and data-labeling costs in semi-supervised setups.

PYTORCH

Portfolio level carbon emissions reduction tool

This study explores the application of the FixMatch algorithm to lightweight image classifiers, aiming to optimize semi-supervised learning for resource-constrained environments. Leveraging the CIFAR-10 dataset, the study evaluates the performance of three architectures: ResNet-8, ResNet-20, and a custom OptVGGNet model. The findings highlight the potential of lightweight models in achieving efficient learning while reducing computational demands and data-labeling costs in semi-supervised setups.

PYTORCH

Software Stack

Software Stack

Python

Library

Python

Library

Python

Library

Python

Library

Python

Library

Python

Library

© Alper Erten. 2025

CREATED WITH LOVE IN SAN FRANCISCO