Ferhat Yaman

Ferhat Yaman

Security Researcher at AMD | Expert in Side-Channel Analysis, Post-Quantum Cryptography & Homomorphic Encryption | Advancing Secure Machine Learning

I am a Hardware Security Researcher at AMD, specializing in secure cryptographic hardware architectures and software systems to safeguard the privacy and integrity of personal computers, cloud infrastructures, and machine learning accelerators. I earned my M.Sc. in Computer Engineering from North Carolina State University, where I was co-advised by Dr. Samira Mirbagher Ajorpaz and Aydin Aysu. My thesis, titled Agent SCA: Advanced Physical Side Channel Analysis Agent with LLMs, focused on leveraging large language models for advanced physical side-channel analysis. Previously, I completed a double major in B.Sc. Computer Science & Engineering and B.Sc. Electronics Engineering at Sabancı University.

My research interests include side-channel analysis, post-quantum cryptography, homomorphic encryption, and privacy-enhancing technologies.

Publications

  1. Yaman, F., Mert, A.C., Ozturk, E., Savas E., “A Hardware Accelerator for Polynomial Multiplication Operation of CRYSTALS-KYBER PQC Scheme” Design, Automation & Test in Europe Conference & Exhibition (DATE’21)
  2. Kurian, A., Dubey, A., Yaman, F., Aysu, A. “TPUXtract: An Exhaustive Hyperparameter Extraction Framework”. IACR Transactions on Cryptographic Hardware and Embedded Systems, (CHES 2025)
  3. Magara, S.S., Yildirim, C., Yaman, F., Dilekoglu B., Tutas, F.R., Ozturk, E., Kaya, K., Taştan, O., and E Savas, E., “ML with HE: Privacy Preserving Machine Learning Inferences for Genome Studies”. ACM Computer and Communications Security, CCS’21, PPML Workshop
  4. Calhoun, A., Ortega, E., Yaman, F., Dubey A., Aysu, A., “Hands-On Teaching of Hardware Security for Machine Learning” Proceedings of the Great Lakes Symposium on VLSI, (ACM GLS-VLSI 2022)
  5. Mert, A.C., Yaman, F., Karabulut, E., Ozturk, E., Savas, E., Aysu, A., “A Survey of Software Implementations for the Number Theoretic Transform” SAMOS 2023 International Conference on Embedded Computer Systems

Work Experiences

AMD

Hardware Security Reseacher – Product Security Research Team (PSO RD), May 2022 – Present

  • Conducted in-depth evaluations of Post-Quantum Cryptography (PQC) algorithms, ensuring optimal efficiency and robustness for next-generation cryptographic systems.
  • Architected secure computation frameworks for machine learning (ML) and artificial intelligence (AI) on GPUs, including the design of a homomorphic encryption compiler to protect sensitive AI operations.
  • Strengthened the security architecture of AMD’s Secure Encrypted Virtualization and Trusted Execution Environment, mitigating potential vulnerabilities and enhancing system integrity.
  • Designed and implemented test environments to validate Secure Memory Encryption and cryptographic components for system-on-chip (SoC) platforms, contributing to FIPS certification readiness.
  • Enhanced the resilience of cryptographic coprocessors against side-channel attacks (SCA) by developing advanced models using ChipWhisperer Python libraries and creating custom SCA models using Riscure Inspector, driving innovative approaches to secure cryptographic processors and mitigate emerging threats.

Accenture

Software Engineering Intern, Jun 2019 – Jul 2019

  • Contributed to the development of a human resources automation system, streamlining routine HR tasks and enhancing operational efficiency.
  • Designed and implemented RESTful APIs for backend services, enabling seamless data exchange and improving system functionality. Integrated event-driven and time-triggered serverless applications into a microservices architecture, leveraging Microsoft Azure to ensure scalability and reliability.

Vingd

AI Software Engineering Intern, Jan 2019 – Feb 2019

  • Designed and implemented a rule-based security AI bot using Bayesian Networks in Java, and deployed the solution in a containerized environment using Docker for scalability and reliability.

Multinet inventiv

Software Engineer, Jul 2018 – Jan 2019

  • Developed a Fraud Detection and Prevention System leveraging outlier detection machine learning models to identify anomalies in large-scale payment datasets using Weka.
  • Designed and implemented database schemas using Java Hibernate and Spring Framework, enabling efficient data management and integration with backend services.
  • Built and tested RESTful APIs for machine learning model deployment and system functionality, ensuring high reliability through JUnit-based validation.

Research Experiences

Hardware Cyber Threat Research Lab, NC State University

Advisor: Dr. Aydin Aysu, Jan 2021 – July 2022

  • Conducted side-channel analysis (SCA) on deep learning accelerators for edge devices, leveraging the Python TensorFlow framework to uncover vulnerabilities.
  • Designed and implemented attack vectors targeting Tensor Processing Units using Riscure Inspector, enhancing the understanding of potential security flaws.
  • Optimized solvers for lattice-based cryptography problems, including NP-hard challenges like the Shortest Vector Problem (SVP) and Learning with Errors (LWE), contributing to advancements in PQC

Computer and Information Security Lab, Sabanci University

Advisors: Prof. Erkay Savas, Prof. Albert Levi, Sept 2019 – Jan 2021

  • Optimized Homomorphic Encryption (HE) and Post-Quantum Cryptography schemes, contributing to advancements in secure computing.
  • Developed Python simulations for prototyped FPGA designs, enhancing performance and efficiency.
  • Built privacy-preserving machine learning applications using HE in a Python-based DNN framework.
  • Implemented a secure inference system combining homomorphic encryption and machine learning, securing a top-5 finish in the IDASH’20 competition.
  • Applied machine learning models to network traffic data for intrusion detection, improving security in IoT.

Institute for Infocomm Research, A-STAR

Advisors: Dr. Chao Jin, Dr. Ahmad Al Badawi, July 2019 – Sept 2019

  • Developed and implemented privacy-preserving protocols for Deep Learning models, enhancing data security and model performance.
  • Leveraged expertise in Multiparty Computing Security (Homomorphic Encryption, Garbled Circuits) to optimize CNN models, reducing resource usage while maintaining efficiency and security.

Secure Systems Lab, Boston University

Advisors: Dr. Manuel Egele, July 2017 – Sept 2017

  • Conducted research and analysis to identify and mitigate potential security vulnerabilities in digital platforms.
  • Analyzed Twitter data for homoglyph character vulnerabilities using Python, identifying potential risks in user interactions. Developed a scalable web scraper system utilizing a 30-node cluster and RabbitMQ