Cleveland Crossling
builds intelligent systems.

AI engineer and data scientist currently pursuing an MSc in Computer Science (AI & Machine Learning) at East Tennessee State University, where I work as a Research Assistant on autonomous vehicle driving models. I design and ship machine learning systems from medical imaging models to visual neuroevolution and autonomous driving systems.

scroll

The short version

About

Computer Science graduate (BScHons, Cum Laude, Top Project Award) now pursuing an MSc in Computer Science with a concentration in Artificial Intelligence and Machine Learning at East Tennessee State University, where I work as a Research Assistant on autonomous vehicle driving models. My work spans applied AI, computer vision, NLP and graph-based modelling, with international research exposure along the way.

I like taking intelligent systems from a research idea to something that actually ships, across healthcare, sports analytics, and autonomous systems. Seeking opportunities in data science and applied AI research where I can apply machine learning and statistical modelling to real-world problems.

Cum Laude
BScHons Computer Science (AI)
MSc @ ETSU
AI & Machine Learning concentration
Nagoya, JP
Published & presented research
& more
See the full CV ↗

What I work with

Skills

Languages

PythonJavaJavaScriptC#C++SQLx86 Assembly

Tools & Frameworks

PyTorchTensorFlowKerasScikit-learnPandasNumPyMatplotlibGit & GitHubASP.NETMS SQL Server

Data Science & AI

Feature EngineeringEDANeural NetworksAutoencodersNLPComputer VisionClusteringGraph ModellingTransformersNeuroevolution

Some of my work

Featured Projects

Personalized PyTorch Frameworks

Custom data preparation, dataloader, training, and evaluation frameworks for streamlining new ML research projects.

MLOpsPyTorch
◆ New · Interactive demo

From-Scratch I-JEPA + Token Explorer

A self-supervised model built by hand in CuPy, now training end-to-end on MNIST. Click through the patch tokens it learned, their attention, nearest neighbours and t-SNE.

Self-SupervisedI-JEPA
Launch the explorer →

Brain Tumor Segmentation

Efficient transformer variants in a 3D V-Net, tackling both accuracy and compute constraints for Sub-Saharan deployment.

Computer VisionPyTorch

Talks & demos

Videos

A talk exploring some of the biggest open questions and unresolved challenges in artificial intelligence today.
A breakdown of Receptance Weighted Key Value (RWKV), a recurrent alternative to transformer attention.
A demo of a Q learning agent navigating mazes generated with depth first search.
A presentation on using convolutional neural networks to detect and classify invasive wildlife from trail camera footage.

Let's talk

Get In Touch

Open to data science and AI roles, reach out if you'd like to talk.

Currently based in the United States