Researcher | Trustworthy AI

SnT

Biography

I defended my Ph.D. at FCUP in Computer Science with a research focus on data privacy and utility preservation of machine learning models. Currently, my work centers on privacy in AI: developing and evaluating methods that enable the responsible use of AI, uncovering how models can leak sensitive information, and designing defences that preserve their utility. Passionate about trustworthy AI, I’m committed to advancing technologies that safeguard individuals' sensitive information.

Education

  • PhD in Computer Science, 2025

    Faculty of Sciences, University of Porto

  • MSc in Network and Information Systems Engineering, 2019

    Faculty of Sciences, University of Porto

Experience

 
 
 
 
 
Postdoc Researcher
Interdisciplinary Centre for Security, Reliability and Trust (SnT)
Apr 2025 – Present Luxembourg

Research topics include:

  • Trustworthiness in AI
  • Counterfactual Explanations
  • Foundation Models
  • Agentic AI
 
 
 
 
 
Data Science Researcher
TekPrivacy
Feb 2020 – Apr 2025 Porto

Research topics:

  • Federated Learning
  • Privacy Enhancing Technologies (PETs)
  • Privacy-Utility Trade-off
  • Meta-learning
  • LLMs
 
 
 
 
 
Visiting Researcher
Lucy Family Institute for Data and Society | University of Notre Dame
Mar 2023 – Jun 2023 Indiana, United States

Developed projects:

 
 
 
 
 
Data Scientist Trainee
NOS
Oct 2018 – Sep 2019 Porto

MSc thesis - Antecipation of Services Perturbance

  • Classification tasks
  • Imbalance domains
  • Sliding windows
  • Big data

Interests

Machine Learning

  * Tabular Foundation Models
  * Agentic AI
  * Generative AI
  * Federated Learning
  * Meta-learning

Open Data

  * EDA
  * Record Linkage
  * Privacy-Utility trade off
  * Secure Data Sharing

Data Privacy

* De-identification
* Privacy-preserving techniques
* Synthetic Data
* Privacy risks

Accomplish­ments

Best Paper Award
22nd Portuguese Conference on Artificial Intelligence (EPIA 2023) Sep 2023
With the paper A Three-Way Knot: Privacy, Fairness, and Predictive Performance Dynamics

Projects

Attention-based MIA and target label k-anonymity defence

Attention-based MIA and target label k-anonymity defence

Membership inference attack in tabular foundation models and inference-time defence.

Follow
ε-PrivateSMOTE

ε-PrivateSMOTE

Differentially-Private Data Synthetisation for Efficient Re-Identification Risk Control

Follow
Hands-on data de-identification

Hands-on data de-identification

Practical course covering the basic principles of the data de-identification process.

Follow

Publications

Tânia Carvalho, Nuno Moniz, Luís Antunes (2025). Automated Privacy-Preserving Techniques via Meta-Learning. In Information Fusion (Explainable AI in Industry 4.0 and 5.0).

PDF Cite Code

Tânia Carvalho, Nuno Moniz, Luís Antunes, Nitesh Chawla (2025). Differentially-private data synthetisation for efficient re-identification risk control. In Machine Learning.

PDF Cite Code

Pedro Santos, Tânia Carvalho, Filipe Magalhães, Luís Antunes (2025). Secure Visual Data Processing via Federated Learning. In ICISSP.

PDF Cite

See all publications

Collaborations / Activities

Track chair
I have been part of the organising committee for the Ethics and Responsibility in AI thematic track at the Portuguese Conference on Artificial Intelligence (EPIA) since 2025.
Feb 1, 2025
EPIA 2025
Track chair
Business Mentor
I was a mentor at the 7th Hack a City event, where I provided guidance and support to participants to improve the city of Porto’s data safely. I was also responsible for the prior de-identification of such data.
Nov 1, 2024
Hackacity
Business Mentor
Collaborator
I explored the use of LLMs to assist in the creation of information sheets and informed consent forms for clinical trials. This study was conducted within the context of the research project REPO4EU (Precision drug REPurpOsing For EUrope and the world).
Jun 1, 2024
RePo4EU
Collaborator

Supervision

Alumni

2024–2025

  • Cristina Pêra (MSc) - Membership Inference Attacks: Evaluation and Defenses
  • Tiago Eusébio (MSc) - Automatic Detection of Personal Information in Tabular Data

2023–2024

  • Miguel Ramos (MSc) - Advancements in Vertical Federated Learning: Utility Preservation with a Lightweight Model
  • Pedro Santos (MSc) - AI-Powered visual Privacy and Labeling

2022–2023

  • Carolina Trindade (MSc) - Identity Disclosure in Synthetic Data
  • Gustavo Pereira (MSc) - Automated Algorithm for Privacy Preservation of Data

Contact