I'm a machine learning researcher and engineer with a PhD in Computer Science from the University of Memphis (August 2025). My research focuses on privacy-preserving federated learning and adversarial ML defense. I have developed algorithms that achieve 90% accuracy while blocking 80% of backdoor and inference attacks in distributed learning systems.

I specialize in building secure and robust ML systems at scale and translating research into production-ready solutions.

• 🎓 PhD researcher in federated learning focused on privacy-preserving ML
• 🔬 Currently building collaborative training algorithms that keep data decentralized
• 📖 Learning advanced differential privacy, secure aggregation, and edge computing
• 🤝 Open to collaborate on research projects, federated learning frameworks, and privacy audits
• 💬 Ask me about privacy-preserving AI, distributed systems, reproducible research, and academic writing
• 📫 Reach me: Email • LinkedIn • Google Scholar

• PhD in Computer Science - University of Memphis (August 2025)

  • Dissertation: "Towards Trustworthy Federated Learning: Enhancing Robustness, Privacy, and Reliability in Collaborative AI"
  • Developed FedTruth and PriFedTruth algorithms for Byzantine-robust and privacy-preserving federated learning
  • Published at top-tier conferences: ICDCS '24, GLOBECOM '24, ACISP '25

• Graduate Cybersecurity Intern - Lawrence Livermore National Laboratory

  • Built PyTorch-based graph neural network pipelines for binary function security analysis
  • Implemented SBOM integration for supply chain security and vulnerability detection

• Master of Science in Computer Science - University of Memphis (2020)

  • Focus: Cloud security architecture, cryptography, and attribute-based access control systems

Languages: Python, C++, Java, SQL

ML/AI: PyTorch, TensorFlow, scikit-learn, Graph Neural Networks, Transformers, Federated Learning frameworks

Security & Privacy: Adversarial ML, Homomorphic Encryption (CKKS), Cryptographic protocols, Backdoor attack defense

MLOps/Infrastructure: Docker, Kubernetes, AWS (EC2, S3, SageMaker), Git, CI/CD pipelines

Distributed Systems: Federated Learning, Distributed training, Multi-party computation

• Building portfolio projects demonstrating end-to-end ML engineering and security expertise
• Exploring LLM applications with privacy-preserving techniques
• Sharpening algorithmic problem-solving skills for production ML systems
• Contributing to open-source ML security tools

• 2025 - Identifying the Truth of Global Model: A Generic Solution to Defend Against Byzantine and Backdoor Attacks in Federated Learning - Information Security and Privacy (ACISP 2025, LNCS) - Springer
• 2025 - Towards Trustworthy Federated Learning: Enhancing Robustness, Privacy, and Reliability in Collaborative AI - PhD Dissertation, Department of Computer Science, University of Memphis - ProQuest
• 2024 - Ensuring Fairness in Federated Learning Services: Innovative Approaches to Client Selection, Scheduling, and Rewards - 44th IEEE International Conference on Distributed Computing Systems (ICDCS 2024) - IEEE CSDL
• 2024 - FedTruth: Byzantine-robust and Privacy-preserving Federated Learning - IEEE ICDCS 2024
• 2024 - Towards Fair, Robust and Efficient Client Contribution Evaluation in Federated Learning - IEEE Global Communications Conference (GLOBECOM 2024) - arXiv
• 2023 - Multi-Criteria Client Selection and Scheduling with Fairness Guarantee for Federated Learning Service - arXiv preprint (CoRR) - arXiv
• Full publication list on Google Scholar

FedTruth Research Implementation - Open-source Byzantine-robust federated learning algorithm achieving 90% accuracy while defending against 80% of backdoor attacks
[Repository link]

Custom GPT Language Model - 6-layer transformer (10–15M params) trained on CUDA; reduced training time by 80% with a modular AWS deployment
[Repository link]

GNN Binary Function Security Analysis - PyTorch-based pipeline across ~50k LoC of firmware; improved detection speed and accuracy for high-risk functions
[Repository link]

SBOM Integration for Software Assurance - MongoDB-backed CVE tracking with ML heuristics; surfaced ~115 hidden discrepancies in supply chain scans
[Repository link]

Homomorphic Encryption in FL - Integrated CKKS into federated learning pipelines, preventing inference attacks across three benchmark datasets
[Repository link]

• LinkedIn: linkedin.com/in/sheldon-ebron
• Email: [email protected]
• Google Scholar: Google Scholar

💼 Open to ML Engineer, Research Scientist, AI Security, and Applied AI roles