👋 Hello, World!
Angelo
Saadeh
Data Privacy · Data Quality · Machine Learning
I am a research fellow within the Valda team of the Computer Science Department at École Normale Supérieure – PSL, in the framework of the PR[AI]RIE Artificial Intelligence Cluster.

Valda is a joint team between Inria, CNRS and ENS-PSL. More information about Valda can be found here.

I currently develop algorithms for data provenance, uncertainty, and probabilistic inference in SQL systems. Previously, I was a research fellow at CNRS@CREATE in Singapore (Descartes program, WP2: Hybrid AI), where I built voting algorithms for ensemble learning and added differential privacy to machine learning pipelines.

Before that, during my Ph.D. at Télécom Paris, I worked on privacy-preserving ML systems using cryptographic tools and differential privacy. 😎

Get in Touch

Have a question or want to collaborate? Send me a message below.

Resume
Experience & Education
Experience
Researcher
École Normale Supérieure, Paris
  • Implement provenance and uncertainty features in ProvSQL (advanced SQL, semirings).
  • Develop efficient algorithms for provenance circuits and probabilistic inference.
  • Analyse complexity and semantics of provenance for advanced SQL queries.
Researcher
CNRS@CREATE, Singapore
  • Integrate data privacy mechanisms like differential privacy into aggregation, training, and inference.
  • Design ensemble learning methods for aggregation across multiple heterogeneous models.
  • Build permutation-invariant machine learning models to measure the quality of data.
  • Construct integer-based mappings that linearize multidimensional data spaces for smart indexing.
Graduate Researcher
École Polytechnique, France
  • Construct numerical methods algorithms to simplify complex functions.
Graduate Researcher
Télécom Paris, France
  • Design hybrid privacy-preserving machine learning models with cryptographic tools and differential privacy.
  • Implement secure federated learning protocols ensuring data confidentiality across multiple parties.
  • Deploy and test a socket network for secure two-party computations in machine learning tasks.
Research Intern
CNRS@CREATE, Singapore
  • Apply secure computation protocols for unsupervised information retrieval algorithms; truth discovery.
Research Intern
Inria Saclay, France
  • Accelerate computations on encrypted data using mathematical tools like error-correcting codes and hyperinvertible matrices.

Education
Ph.D. in Computer Science
Télécom Paris, France
Nov 2019 – Jun 2023
Thesis: Applications of Secure Multi-party Computation Machine Learning
Advisors: Daniel Augot (Inria) and Matthieu Rambaud (Télécom Paris)
Doctoral scholarship: Labex Digicosme
M.Sc. in Computer Science
Université Paris-Saclay, France
Sep 2017 – Sep 2019
Specialization: Cryptography and Applied Algebra
Thesis: Mathematical Tools for Secure Multi-party Computation
B.Sc. in Mathematics
Lebanese University, Lebanon
Sep 2014 – Jul 2017

Skills
Programming
PythonCC++Java
Libraries & Tools
PyTorch / PySyftTensorFlow / TF-encrypted PandasScikit-learn Socket NetworksLaTeX
Databases & Web
SQL / PostgreSQLXML / XQueryHTML
Languages
English (fluent)French (fluent) Arabic (fluent)Italian (B2) Mandarin (HSK2)
Research
Publications & Projects
Discovering Voting Power for Ensemble Methods
DEXA 2025 · Bangkok, Thailand · Lecture Notes in Computer Science, Vol. 16046
Pratik Karmakar, Angelo Saadeh, Pierre Senellart, and Stéphane Bressan
doi →
TF-MPC: Confidential Truth-Finding with Multi-Party Computation
GitHub Repository · 2022
Angelo Saadeh
github →
Applications of Secure Multi-party Computation in Machine Learning
PhD Thesis · Institut Polytechnique de Paris · 2023
Angelo Saadeh
hal →
Epsilon-Differentially Private and Fully Secure Logistic Regression on Vertically Split Data
ICDIS 2022 · 4th International Conference on Data Intelligence and Security
Angelo Saadeh, Vaibhavi Kumari, and Stéphane Bressan
doi →
Confidential Truth Finding with Multi-Party Computation
DEXA 2023 · Penang, Malaysia · Springer
Angelo Saadeh, Pierre Senellart, and Stéphane Bressan
doi →
WannaFly: Dummy Ransomware for Red Team Exercises
REDOCS 2021 · Airbus
Aboubacar Djibo Maman Sani, Boukerrou Hamid, Marinho Dylan, Saadeh Angelo, and Somers Benjamin
pdf →
Teaching
Courses & Lab Sessions
INFMDI731
Advanced Cryptography
Télécom Paris · Spring 2021
Intervention: Secure Multi-Party Computation
MDI210
Optimization and Numerical Analysis
Télécom Paris · Lab sessions: Fall 2020  ·  Course: Fall 2021
Course lectures and Java lab sessions covering optimization methods and numerical techniques.
BT5110
Data Management and Warehousing
National University of Singapore (NUS) · Lab sessions: Fall 2023
PostgreSQL lab sessions covering data management, warehousing concepts, and SQL querying techniques.
Cool Stuff
Side Projects
Kalendar

A tile puzzle: place 6 out of the 7 pieces on the 7×5 board to cover every cell except today's date. Drag pieces from the tray, rotate & flip with the button. Every day has many solutions, you can see them if you click "Show Solution" to see one possible solution.

Target day:
Board
Pieces
Chinese Flashcards
Practice HSK vocabulary with flashcards. Click the card to reveal pinyin, then click again for the definition. Select multiple HSK levels below. Vocabulary for HSK5 and HSK6 is currently not available, if you want them, contact me and I'll add them for you :-)
⟶ View on GitHub
click to reveal
0/0
HSK-3 Français
Cartes Mémoire · Sélectionner les leçons pour réviser le vocabulaire. Cliquer sur la carte pour afficher le pinyin, puis cliquer à nouveau pour faire apparaître la définition.
cliquez pour révéler
0/0
🔐 Secure Multi-party Computation
Simulate an addition protocol for n parties. All computations are done modulo 220 = 1048576.

Step 1. Choose number of parties (n) and hidden parties (k). You won't have access to the information of the hidden parties

Oh what's that?

A special message just for you!
🗺️ the decision matrix · thailand ✦ japan
🇹🇭 Thailand
bangkok · krabi · phi phi · phuket
⚡ 27–32°C · humid · hot · rainy · tropical
  • 🍜 pad thai · curry · mango sticky · tropical fruits
  • 🏝️ swimming · snorkeling · diving · island hopping · kayaking around cliffs · beach bars and sunset spots
  • 🏙️ city · temples · night markets · rooftop bars · shopping malls · large nightlife scene
  • 💸 street food €2–5 · hotel €40–120
🗾 Japan
tokyo · kyoto · osaka · nara
🍁 5–20°C · autumn · warm · chilly
  • 🍣 sushi · ramen · yakitori (mashewe) · tonkatsu
  • ⛩️ temples · shrines · gardens · bamboo forests · tea houses · anime · samurai · ninja
  • 🏙️ neighborhoods (shibuya | shinjuku) · futuristic · shopping and fashion · good nighlife scene
  • 💴 meal €8–15 · hotel €70–200
📝
🇹🇭 Thailand
✈️ Travel
  • Domestic flights: €30–70
  • Ferries: €10–25 to islands
  • Buses/vans: €3–15
💰 Costs
  • Street food: €1.50–4
  • Casual restaurant: €6–15
  • Nice hotel: €45–130
  • Luxury resorts: €200+
🎒 Travel style
  • Very flexible · last minute
  • Accommodation widely available
  • Excellent backpacker infra
🗾 Japan
🚄 Travel
  • Tokyo → Kyoto: €95–110 · 2h15
  • Kyoto → Osaka: €10–15 · 15min
  • Rail Pass may save money
💰 Costs (weak yen helps)
  • Convenience meal: €5–8
  • Casual ramen: €8–12
  • Dinner with drinks: €20–45
  • Nice hotel: €80–220
  • Hostel dorm: €20–35
📋 Travel style
  • Book hotels in advance
  • Reservations recommended
  • Highly organized · punctual
  • Less spontaneous
✅ thailand pros
  • incredible food · beaches
  • very affordable · easy travel
  • friendly & flexible
⚠️ thailand cons
  • intense humidity
  • rudimentaire / poor pockets
  • 3al lebnene vibe
✅ japan pros
  • extremely safe & clean
  • food culture · autumn
  • tradition + futurism
⚠️ japan cons
  • more expensive
  • less spontaneous
  • language barrier
🤔 still unsure?

jungle city · relaxed tropical trip · beaches · cheap food · spontaneity · warm nights · sunsets → Thailand

cultural journey · temples · incredible cities · autumn colors · unforgettable food → Japan

we will enjoy it no matter what we choose