Add CoDaS-HEP 2026 participant: Marco Zaccheddu

_codas-hep-students/2026/zaccheddu.md

@@ -0,0 +1,30 @@
+---
+layout: codas-hep-participant
+e-mail: zacch@jlab.org
+institution: Jefferson Lab
+name: Marco Zaccheddu
+photo: "/assets/images/codas-hep/2026/Marco-Zaccheddu.png"
+github-username: zaccheddu
+linkedin-profile: www.linkedin.com/in/marco-zaccheddu-2853572a0
+orcid: https://orcid.org/0000-0002-3672-8111
+title: Postdoctoral Researcher
+website:
+logos:
+    - /assets/images/codas-hep/logos/JLab_logo.jpg
+    - /assets/images/codas-hep/logos/Iris-hep-logo.png
+---
+
+## My research:
+I am a postdoctoral researcher in the Theory and Computation Division at Jefferson Lab. My research focuses on hadron physics and structure, with a particular emphasis on hadron spin structure. I specialize in extracting parton distribution functions (PDFs) by leveraging machine learning methods and generative AI.
+
+## My expertise is:
+My expertise lies in TMD (Transverse Momentum Dependent) and GPD (Generalized Parton Distribution) physics. I have extensive experience in solving inverse problems using Bayesian approaches and generative AI techniques, such as Normalizing Flows.
+
+## A problem I'm grappling with:
+Currently, I am focusing on how to properly characterize and improve uncertainty quantification for PDF extractions, particularly when dealing with complex linear and bilinear inverse problems.
+
+## I've got my eyes on:
+I am closely following the adaptation of generative AI architectures, such as Convolutional Neural Networks (CNNs) and Transformers—traditionally used in computer vision and image processing—and exploring how they can be effectively applied to the extraction of PDFs.
+
+## I want to know more about:
+I am eager to learn more about advanced software optimization techniques, accelerating ML workflows using modern GPU architectures, and best practices for scaling generative models within High-Performance Computing (HPC) environments for HEP applications.