Jean-Philippe Bossuat

Jean-Philippe Bossuat

Cryptography Research Scientist & Software Specialist

GAUSS LABS PTE. LTD.

Hi, I am the principal developer and co-author of the library Lattigo, a Go module that implements Ring-Learning-With-Errors-based homomorphic-encryption primitives and Multiparty-Homomorphic-Encryption-based secure protocols.

I have been working as a research scientist and cryptography software engineer for the past five years. My work consists in researching, designing, implementing and integrating privacy preserving machine learning and statistical analytics in large scale B2B applications. The designed solutions are built on top of applied cryptography combined with other privacy enhancing technologies, such as lattice based fully homomorphic encryption, multiparty computation, federated learning and differential privacy.

Interests
  • Privacy Preserving Machine Learning and Analytics
  • Lattice Based Homomorphic Encryption and Secure Multiparty Computation
  • EU/CH Privacy Laws Compliance
Education
  • Two graded semesters at EPFL on privacy enhancing technologies, number theory and applied cryptography, 2019

    École Polytechnique Fédérale de Lausanne (EPFL)

  • Master in Law (MLAW) in Legal Issues, Crime and Security of Information Technologies, 2019

    Université de Lausanne (UNIL)

  • Bachelor in Law (BLAW), 2016

    Université de Lausanne (UNIL)

Experience

 
 
 
 
 
Senior Cryptographer - Homomorphic Encryption
GAUSS LABS PTE. LTD (Singapore)
April 2024 – Present Lausanne - Switzerland
 
 
 
 
 
Tune Insight
Privacy Preserving Machine Learning and Analytics Cryptography Research Scientist & Software Specialist
November 2022 – April 2024 Lausanne - Switzerland

Responsibilities include:

  • Project lead of the cryptographic library Lattigo
  • Research, design, implement and optimize low- and high-level cryptographic primitives, protocols and FHE/MPC-based applications for B2B privacy preserving analytics and machine learning
  • Coordinate and carry out state of the art research and collaboration with external entities in the field of lattice based homomorphic encryption
  • Consolidate Tune Insight’s position within the cryptographic community
  • Represent Tune Insight in the iDASH challenges
 
 
 
 
 
Laboratory for Data Security - EPFL
Cryptography Software Specialist & Research Scientist
Laboratory for Data Security - EPFL
March 2019 – December 2022 Lausanne - Switzerland
  • Co-author and main contributor of the cryptographic library Lattigo
  • Design and optimize applications using fully homomorphic encryption
  • Contribute to the research of the PHD students
  • Collaborate with external entities for privacy law compliance
  • Represent the Laboratory for Data Security in the iDASH challenges
  • Teaching assistant

iDASH

The iDASH Privacy and Security Workshop is an interdisciplinary competition motivated by real world biomedical applications.

Track I Secure Neural Network Evaluation for Protein Sequence Classification
IDASH PRIVACY & SECURITY WORKSHOP 2024 - Secure Genome Analysis Competition
  • Task: Homomorphic evaluation of a transformer to classify encrypted sequences of proteins
  • Ranking: Among the winnng teams, final ranking discolsed the 9.11
  • Team: Jean-Philippe Bossuat (Gauss Labs)
    • A special thanks to Seonhong Min and Intak Hwang for their contribution to GoFHE and insightful discussions during the finalization of the solution
  • Presentation: available the 9.11
  • Code: GitHub
  • Backend: GoFHE
Track I Secure Relative Detection in (Forensic) Databases
IDASH PRIVACY & SECURITY WORKSHOP 2023 - Secure Genome Analysis Competition
Track II Secure Model Evaluation on Homomorphically Encrypted Genotype Data
IDASH PRIVACY & SECURITY WORKSHOP 2022 - Secure Genome Analysis Competition
  • Task: Securly outsource phenotype prediction on encrypted genomes
  • Ranking: 3rd Place
  • Team: Jean-Philippe Bossuat (Tune Insight), Thierry Bossy (EPFL), Juan Troncoso-Pastoriza (Tune Insight)
  • Presentation: YouTube
  • Backend: Lattigo v3.0.6
Track II Homomorphic Encryption based Secure Viral Strain Classification
IDASH PRIVACY & SECURITY WORKSHOP 2021 - Secure Genome Analysis Competition
  • Task: Securely classify encrypted genomes of the SARS-CoV-2 among different strains
  • Ranking: 2nd Place
  • Team: Jean-Philippe Bossuat (EPFL), Sinem Sav (EPFL), Juan Troncoso-Pastoriza (EPFL), Jean-Pierre Hubaux (EPFL)
  • Presentation: YouTube
  • Code: GitHub
  • Backend: Lattigo v2.1.1
Track II Secure Genotype Imputation using Homomorphic Encryption
IDASH PRIVACY & SECURITY WORKSHOP 2019 - Secure Genome Analysis Competition

Projects

*
GoFHE
GoFHE is a fork of Lattigo v5.0.0, that includes an extensively refactored backend, many additional features and new protocols.
GoFHE
Lattigo: lattice-based multiparty homomorphic encryption library in Go
Lattigo: lattice-based multiparty homomorphic encryption library in Go.
Lattigo: lattice-based multiparty homomorphic encryption library in Go

Publications

(2024). Practical q-IND-CPA-D-Secure Approximate Homomorphic Encryption.

Cite Link

(2024). Security Guidelines for Implementing Homomorphic Encryption.

Cite Link

(2023). Scalable and Privacy-Preserving Federated Principal Component Analysis. 2023 IEEE Symposium on Security and Privacy.

Cite PDF Link Trailer Video Video

(2022). Orchestrating Collaborative Cybersecurity: A Secure Framework for Distributed Privacy-Preserving Threat Intelligence Sharing.

PDF Cite DOI

(2022). Privacy-Preserving Federated Recurrent Neural Networks. Accepted for publication at the 23rd Privacy Enhancing Technologies Symposium (PETS 2023).

PDF Cite DOI

Talks

Devcon 2024
Devcon 2024
FHE.org Toronto
Large Domain Homomorphic Evaluation in Levelled Mode
FHE.org Toronto

Teaching

I have been a teaching assistant for the following courses:

Reviews & Sub-Reviews

I have served as a Program Commitee, reviewer or sub-reviewer for the following conferences and journals: