Computer Vision+120k patches processed
Isotropic 3D Bio-Imaging Reconstruction
CycleGAN architecture in PyTorch for unpaired image-to-image translation, correcting anisotropic distortion in tissue microscopy datasets.
Jefferson Conza
Mathematics Student · Machine Learning Engineer
Researching Scientific Machine Learning (PINNs), computer vision, and cloud solutions with AWS. Passionate about history, scientific breakthroughs, and mathematical animation.
Featured Projects
View all projects →Machine Learning99% Sensitivity (Recall)
Pneumonia Detection with Transformers (CheX-DS)
Hybrid DenseNet121 + Swin Transformer in PyTorch with weighted asymmetric loss, achieving 99% recall on chest radiographs.
SciML / PINNsScientific Cloud Deployment
PINNs & SciML Laboratory on AWS
Physics-Informed Neural Networks for PDEs (1D Burgers, 2D thermal diffusion, Navier-Stokes) with FastAPI microservices and AWS EC2 deployment.
Latest Essays
View all essays →OpenAI and the Navier–Stokes Problem: Finite-Time Singularity and the New Mathematical Frontier
An in-depth analysis of OpenAI’s mathematical breakthrough: constructing finite-time blowup solutions for the Navier–Stokes equations in ℝ³ and the periodic torus, the breakdown of unforced Euler, and formal verification in Lean 4.
Graphene: The Carbon Monolayer and Scientific Footprints in Ecuador
From micromechanical Scotch-tape cleavage in Manchester to laboratories at Yachay Tech and ESPOL: the physics of the two-dimensional crystal and pioneering research in Ecuador.
Vision Transformers and CNNs: Hybrid Architectures in Radiological Diagnosis
From Röntgen’s X-rays to the CheX-DS hybrid architecture: how the convergence of dense convolutions and shifted-window self-attention redefines medical imaging.