─── CURRICULUM VITAE ───

JOSÉCABRERO-HOLGUERAS

Senior Software Engineer (Cryptography, Privacy & AI)

Madrid, Spain·PhD Cum Laude·Nillion·AI & Privacy
JOSÉ CABRERO-HOLGUERAS
MAIL·GITHUB·LINKEDIN

OBJECTIVE

Senior software engineer with 7+ years of experience building AI production systems and privacy-enhancing infrastructure. Led development of GPU-accelerated MPC inference engines, TEE-based LLM platforms, and decentralized attestation networks, with published research alongside Meta AI and CERN. Ph.D. (Cum Laude) with peer-reviewed publications in top-tier venues.

EXPERIENCE

Nov. 2023Present

Senior Software Engineer (Cryptography, Privacy & AI)

Nillion—Madrid, Spain
  • ›Confidential AI Platforms: Architected and led development of nilAI, a production TEE-based LLM Inference platform exposing secure multi-tenant APIs with integrated authentication, attestation, and rate limiting; contributed to nilCC for general-purpose confidential compute on AMD SEV-SNP and NVIDIA Confidential Computing.
  • ›Privacy-Preserving Inference: Led development of AIVM, a Meta CrypTen fork for hybrid GPU-accelerated MPC+TEE inference, achieving ~0.1s latency for selected models. Co-authored Fission with Meta AI.
  • ›Decentralized Attestation Infrastructure: Co-led the design and implementation of Blacklight Network, a decentralized TEE attestation verification protocol governed by Solidity contracts, securing $260K+ in total stake.
  • ›Privacy-Preserving Authentication & Payments: Designed and integrated privacy-preserving authentication and decentralized Web3-native payment flows into confidential compute infrastructure.
TEE / Confidential ComputingLLM InferenceMPCSolidityAMD SEV-SNPNVIDIA Confidential Computing
Jul. 2023Nov. 2023

Cryptography Engineer

CSIC—Madrid, Spain
  • ›Development of New Features in Group Signature Libraries: New advances in IBM libgroupsig group signature libraries, focusing on the creation and implementation of new features.
  • ›Integration and Migration of Smart Contracts (dApps) to Hyperledger Besu: Integration of smart contract functionalities with Hyperledger Besu, enhancing the efficiency and effectiveness of blockchain-based transactions.
  • ›Post-Quantum Secure Environments and Communications: Initiatives to establish secure digital environments using post-quantum techniques, ensuring robust protection for different technologies.
Group SignaturesHyperledger BesuPost-Quantum CryptographySmart Contracts
Feb. 2023Apr. 2023

Privacy-Preserving Machine Learning Researcher

Universidad Carlos III de Madrid—Madrid, Spain
  • ›Symbolic Execution Homomorphic Encryption Compiler: Developed an innovative homomorphic encryption compiler using symbolic execution techniques that achieved a remarkable 80% reduction in code lines required for privacy-preserving Deep Learning Inference.
Homomorphic EncryptionSymbolic ExecutionDeep Learning
Nov. 2019Nov. 2022

Privacy-Preserving AI Researcher — Doctoral Student

CERN—Geneva, Switzerland
  • ›Researching on Privacy-Preserving Methodologies for Deep Learning. Application of Homomorphic Encryption to Deep Learning models for its privatization. Easing access for programmers for modern cryptographic techniques, including parameter selection and vectorization of Deep Learning operations.
  • ›Integrated expert systems, leveraging fuzzy logic and linear programming techniques to automatically generate optimized parameters for homomorphic encryption.
  • ›Designed novel algorithms specifically tailored for convolutional neural networks with packed homomorphic encryption that achieved a 2× performance improvement from existing ones.
  • ›Research Management: Took a main role in scientific writing and supervision and guidance to other students and interns. Actively contributed to new research project initiatives, discussions, and writing at a European level.
Homomorphic EncryptionDeep LearningSymbolic ExecutionResearch Management
Jun. 2018Aug. 2018

Embedded Systems Software Intern

CERN, Detector Technologies Department—Geneva, Switzerland
  • ›Embedded Readout and Management System for Radiation Sensors: Developed a minimal platform for data extraction, database storage, and result visualization on an Arduino Yun, achieving a minimal memory footprint of 64 MB.
Embedded SystemsArduinoData Visualization
Sep. 2017Jun. 2018

Compiler Development Intern

ARCOS group, Universidad Carlos III de Madrid—Madrid, Spain
  • ›Support for C++ Contract-Based Programming on the Clang compiler: Integrated contract-based programming into Clang compiler, enabling enforcement of preconditions, postconditions, and assertions in C++ code, resulting in up to 40% performance improvement.
C++ClangCompiler Development