Julien Delaunay

Julien DELAUNAY

PhD Researcher in Computer Science

Top Doctors, Barcelona

About Me

I am a researcher and engineer dedicated to the intersection of Explainable AI (XAI), Natural Language Processing (NLP), and Human-Computer Interaction (HCI). I earned my PhD from Inria Rennes, where my research focused on the design and evaluation of explanation techniques to enhance user comprehension and confidence in complex machine learning models. My doctoral work specifically explored how the representation and technique of explanations affect human trust and decision-making in high-stakes environments.

In my current role as an R&D Scientist at Top Doctors in Barcelona, I lead the integration of Large Language Models (LLMs) into clinical workflows. My work involves developing robust systems for clinical data structured extraction, automated diagnostic assistance, and patient-facing AI tools that prioritize transparency and reliability. I bridge the gap between cutting-edge research and production-ready healthcare applications, ensuring that AI solutions are both technically sound and ethically grounded.

Beyond technical implementation, I am deeply interested in how humans perceive and interact with AI. I strive to build systems that don't just provide answers, but offer meaningful insights, ensuring that AI remains a supportive and understandable tool for both professionals and the general public. Whether it's through multi-modal interfaces or innovative explanation frameworks, my goal is to create AI that empowers users rather than obscures the truth.

Work Experience

Researcher in NLP & AI @ Top Doctors

2024 - Present | Barcelona, Spain

Developing chatbots to empower patients in specialist discovery and providing intuitive AI explanations for healthcare professionals. Leading LLM integration for clinical workflows.

PhD Student in Explainable AI @ Inria

2020 - 2024 | Rennes, France

Supervised by Christine LargouΓ«t and Luis Galarraga. Researched user comprehension and trust in AI. Developed methods for linear and global explainability in NLP models.

Visiting Researcher @ Aalborg University

2023 | Aalborg, Denmark

Collaborated on evaluating human-centered XAI with Niels van Berkel and the HCI group.

Research Intern @ Inria Rennes / LIRIS

2019 - 2020 | France

Focused on mining intuitive referring expressions in knowledge graph.

Selected Publications

Loading publications... πŸ“‘

View Full Publication List β†’

Teaching

Prediction Methods (2020–2023)

M1 Level – ISTIC, Rennes 1 University

Taught time series analysis (AR models, exponential smoothing) and supervised house price prediction competitions.

Machine Learning & Pandas (2023)

M1 Level – Institut Agro Rennes

Hands-on Python workshops using scikit-learn, numpy, and pandas for real-world data science projects.

Introduction to Neural Networks (2021–2022)

M2 Level – ENSAI

Deep learning fundamentals for tabular and image data, including CV projects like dog vs. cat classification.

Technology Monitoring (2020–2022)

M2 Level – ISTIC, Rennes 1 University

Evaluated student reports and oral presentations on emerging technologies and scientific trends.

Office Tools for Data Science (2020–2022)

L3 Level – ENSAI

Specialized Excel and LaTeX training for professional statisticians across multiple groups.

Object Programming in Java (2020–2022)

L2 Level – ISTIC, Rennes 1 University

Core OOP concepts, inheritance, and data structures. Developed a Tower Defense game as a final project.

πŸ€– Featured Projects

A selection of my work in Medical AI, Explainable Machine Learning, and Data Infrastructure.

NanoRank

NanoRank: Esports Scouting Application

An application for scouting players in competitive games like Rocket League, League of Legends, and Counter-Strike. Built for the Nanocorp ecosystem to analyze player performance and potential.

TDListener

TDListener: A Medical AI Scribe

A tool that leverages AI to automatically transcribe and summarize medical consultations, freeing up doctors to focus on their patients. It processes clinical audio into structured SOAP notes.

Global Data Space

Global: Shared Data Space for Clinical and Hospitals

A secure and collaborative platform for sharing clinical and hospital data, designed to accelerate research and improve patient care through interoperable FHIR standards.

Therapy XAI

Therapy: Global Explainability of NLP Models

This project introduces a method for generating global explanations for NLP models through a cooperative generation process. It helps in understanding the overall behavior of complex models.

APE Research

APE: When Should We Use Linear Explanations?

Investigates the suitability of linear explanation models for different ML models and data instances, providing practical guidance for model interpretability.

Anchors Research

Improving Anchor-based Explanations

Enhancing Anchor-based XAI by improving tabular data discretization and extending search space for more robust textual explanations.

Contact

πŸ“ Barcelona, Spain

πŸ“§ juliendelaunay35000@gmail.com

πŸ”— LinkedIn Profile

πŸ’» github.com/j2launay

Education

PhD Computer Science

Inria Rennes - Rennes 1 University (2020 - 2023)

XAI, NLP, User Studies. Defended Dec 2023.

MS Computer Science

Rennes 1 University (2019 - 2020)

MS Computer Science

Sherbrooke University, Canada (2018 - 2019)

Supervision

2020

Mentoring of Jacques Lacourt, a final year trainee at Centrale Marseille.

Organisation Member

2021 & 2022

Student volunteer for the BDA conference (Rennes, France).

2021 & 2022

Member of the Organisation committee of the PhD student day of the IRISA laboratory (Rennes, France).

2021-2022

Data Knowledge Management Seminars (Inria).

2020-2023

Centre Committee Representative (Inria).

Skills

AI & Machine Learning

XAI, NLP, LLMs, Scikit-learn, PyTorch

Development

Python, JavaScript, SQL, Git

Languages

FR French (Native)

EN English (Fluent)

ES Spanish (Professional)

βœ‰οΈ

Get In Touch

I'm always open to discussing new projects or research collaborations.

πŸ‘€
πŸ“§
πŸ’¬